#1 - Vitalik on Starting New Countries, Upgrading Ethereum, and Improving Yourself

Balaji Srinivasan 00:00

This is the first episode of the network state podcast. I've been on many podcasts. I guess they've had millions of views collectively. And I've probably had millions of people asking me to whenever you can do your own podcast, well, we're doing it. Okay. This podcast is about network states, of course, but it's also about everything we're interested in from the positive vision of human improvement, to you know, obviously cryptocurrency but also by technology, drones, nuclear power, and all the policies we need to actually legalize all this amazing stuff. And we're gonna be talking to tech people. I'm going to be talking to policy, people are going to be talking to interesting people, and giving my own commentary and hopefully, injecting some levity occasionally into the proceedings. Today, I'm actually very happy to have our first guest here, my friend Vitalik weiteren, the founder of Aetherium. Here to talk to us today about Ethereum about network states and starting new countries, and also how to improve yourself as well as his particular life hack. One weird trick that allowed him to learn Chinese in how long was it photonic?

Vitalik Buterin 01:06

I built like three years before I got comfortable enough to talk to people. Yeah,

Balaji Srinivasan 01:10

there you go. One weird trick to learn Chinese in three years of hard work. So he'll tell you what he did. All right. Thank you very much. All right. Well, we're underway. This is the first network seat podcast. And obviously, you know, we've talked a lot about a bunch of things. Like rather than doing you know, tell us about yourself and background, all the stuff. I will assume that that's on tons of other podcasts, I think what is maybe a little topical that we can just talk about is Ethereum from concept to the present day, because we just had the merge. And you know, what is Ethereum? After the merge? What's the future? And you know, then what happens after that? Why don't you tell us about that?

Vitalik Buterin 01:49

Sure. So the Etherium project started around the end of 2013. So what I had been doing before then was, I basically finished a well not really finished, but I was about five months into a Bitcoin trip around the world. This is basically something I had started in June of that year, just visiting what was the Bitcoin community at the time, so I visited like the Libertarians in New Hampshire, the left leaning anarchists in Spain, they Yeah, no more, kind of nor me, Bitcoin community in Switzerland, people in Israel who are working on math and covered coins and zero knowledge proofs and master coin, I had accumulated all of these different ideas and a lot of interesting recollections of various projects that people were working on. And what was interesting in Israel in particular, right, was there was this really high density of people who were excited about this problem of, well, what's next? Right? So we have Bitcoin for payments, what is blockchain for, you know, finance, domain name systems, and more general applications going to look like in general, right? And what would a protocol that could actually support those kinds of applications look like? Their first protocol that was trying to do anything vaguely in that range was covered coins, right, which basically just allowed you to access other kinds of assets, and issue other kinds of assets, make sure it's actually the right other kinds of assets on top of the Bitcoin Blockchain. But then, after this, there was a master coin, which was a financial protocol that supported a couple of applications that were called things like, basically things that we would call defi today,

Balaji Srinivasan 03:59

so all the color code stuff was in it was pre 2013 When people still thought all that stuff could run Bitcoin. And then I think we met in late 2013, if I recall, right before

Vitalik Buterin 04:09

he met in San Francisco, actually right after this intro to Israel, right, where I had started actually working on Colored Coins and baster coin and basically right after I had started asking myself the question like, okay, take these protocols as, as they are, what kinds of tweaks can you make to them to try to make them support more things without making the protocols even bigger, right. So I think at the time I coins and was using the service was Army Knife protocols. And what I meant by that rate was something like Mastercoin. It supported 10 different types of applications, but it just had one transaction type for each application. We know there was a transaction type for like registered domain there was a transaction type for open binary financial contract there is no AGI transactions. type four make a bet on an events with this third party as the arbitrator. And I started with poking at these features to try to save like, for example, oh, what happens if we collapse these three features into one feature? That's like one feature, and it's a little bit more complex, but it supports those applications, and it supports even more applications. The rich tearfulness. Exactly. So Rich statefulness actually, actually came near the end, right. So at first, I was just making these tiny tweaks, then I realized, well, what if we made Mastercoin that had a future for generic two party financial contracts where you can specify a formula? And then after that, finally, what happens if you go beyond to party? And you say, Okay, well, what does an N party contract look like? And that's when we had are these big insights like, well, maybe it makes sense for contracts actually be at accounts, what if it makes sense to, to make it possible to send coins directly to a contract. So a contract is like an actual, like virtual agents that you can interact with, you could call that your grad that, you know, send coins to do all kinds of things with. And that's where the ideas for Aetherium started to emerge. And at the beginning, those ideas were phrased in the form of like, well, how do we modify some of these existing systems, but then as the momentum around the projects kind of grew, as more and more people started getting interested in the idea, then it's kind of the ambition quickly grew. And it quickly basically became a project to create a new and independent blockchain. Right. So that's kind of how things started. Mid and I had written this big document back in November of 2013, that I refined a bit since then, but virtually in November, the theory on white paper, right, and I think maybe even when we met for the first time, I might have tried showing it I'm not even

Balaji Srinivasan 07:03

I think there was Viet It was basically the big room where there was a bunch of people there. And actually, several companies came out of it actually, including block stream. Everybody came in that room, and I said, Who are the founders and everybody raise your hands? And I was like, Okay, there's gonna be it's gonna be interesting. And we were there. Like, there were quite a few folks in the early Aetherium. There was that conference in San Francisco a few months later, were I think it was the one where it was a coin summit or something like that in 2014. Is that right? Or at least 2014? is like when the first like, sort of formal Bitcoin Conference like things. And I think around that time, the whole thing was coming together. And I mean, I think the White Paper was out by then or close to up at that. I don't recall exactly when it was public around the time. I think it was held by then. Yeah. And so that's the beginning. Right? And how much of that early roadmap, you know, the frontier homestead metropolis, Randy, how much of that? Did you guys in your view? Did you stick to what did you end up changing? Obviously, you know, getting to proof of steak was much harder than or took took a while, but it's now out there. So what you know, it's impossible to hit something on the first go, and most people would be very happy to have you know, with Etherion, but what's your retro and the whole thing? Yeah,

Vitalik Buterin 08:15

I think what we ended up doing was base basically in the spirit of the original frontier, homestead metropolis, serenity, yes. Separation. Right. So the those four words, by the way, they were from a blog post that was published on the Ethereum blog by I think it would have been a group to like back in early 2015 a bit before Aetherium launched, right. And basically, the intended definitions of those four stages were the frontier was the launch of the chain, but when the chain was launched, it would be launched with like, very explicit Middlemore marketing and like understanding among everyone that this was still a beta chain and wild west of almost exactly Wild West, lots of things could change. You know, this is the frontier you know, the standard things that the bank was people say in every episode, and then homestead was would be a milestone of getting a little bit out of the woods of the Wild West, right? So basically saying like, okay, you know, a chains run for a while and it's survived, it looks fairly stable. So it's okay to consider it a slightly more mature chain, but then obviously, we still keep improving things, metropolis. Actually, this is the one difference. Originally there was a goal that the Ethereum foundation would create something called mist, which was the Ethereum web browser, right? Basically, a web browser designed around the idea of building and using decentralized applications from inside of it, and it was intended to contain things like an app store, and I think there were even ideas near the big Adding of having a delta would manage the App Store, there would be no security oriented features that could only happen because of the browser. There's just a lot of like really fast that they get and advanced ideas. And I think most of that ended up not happening, right instead, you know, we have like browser wallets, so things like meta mask, the brave wallet, and I think more and more new ones just starting to pop up lately. The one thing that ended up not really happening, I think, Mitropoulos ended up being kind of, well, first of all split up into three stages, right? So there's Byzantium, Constantinople, and Istanbul, the Three Forks that happens between 2018 and 2020, right. And, like instable was the bigger one, it added support for pairings, the cryptographic operation that makes a lot of those knowledge proof protocols possible. And then some other features that were enabled by the other two, and then in 2020, EAP, 1559, that was something that was totally not predicted initially. But it ended up actually being the first like really significant change to the Etherium protocol, right, basically, the idea there was, it's a reform to the fee market. So to what, what happens between you sending a transaction and a transaction getting included in a block and the process for like deciding which transactions get included? How much do miners get paid for including those transactions? What are the limits for how many transactions could get included all of those different variables. And basically, it's like, the kind of econom geek explanation is that it moves the market away from being a first price auction, which is no to be any efficient, and it moves it toward being a fixed price system. But one other way of thinking about it is it adds like some temporary select to the block space. So if there is very temporary growth in demand, one block can be a little bit bigger, the next block can be a little bit smaller. And what these things ensure is basically that, if you like users don't have to make this really complicated calculation of am I going to pay a high a very high fee or get a good at immediately? Or am I going to pay a lower fee and like needlessly wait five or 10 minutes. And instead, we would, if you want to know basically gave us as a few market where transactions would reliably get included in the next like one or two blocks, right. And then after that, we had the merge, which I haven't gotten to that quite yet. But I'll kind of poke into the the future of the past here briefly, the merge would obviously switch as to proof of stake. But one of the interesting side effects was that it made block times more regular, right? Instead of a Poisson distribution, where you had like a 1/12 of a chance of a block every second, with a merge, it's like a consistent read or break, block, 12 seconds block four seconds block. And that actually, yeah, meant that because of this extra regularity and block time blocks would happen. On average, you actually have to wait two times less before seeing that extra block, even though the average time between blocks is is exactly the same. It's this really interesting mathematical artifacts, right? And that kind of stacked on top of Yeah, beyond 559, right? Because if you want to place off nine to reduce how long do users have to wait before they get included? And then the words reduce that even further. And so the both of those two things together, like I think the effects of them both were just so visible. The difference is night and day, right? Like sending a transaction and waiting for it to get included back in 2019 was like a very frustrating experience. And today, it's like, click wait, check, ether scan, it's done. Click wait, check, ether scan, it's done. Right. It's like, very often a transaction gets equated within only a few seconds. Like I think probably you get a 50% chance after sending a transaction that it will get included in less than about seven seconds or so. Right? The only reason why it's not six is just the probability that like, whoever the next block creator is just happens to be doing something weird and you have to wait for the next one. So you have some proofs Aetherium clients approved this is also a really important one right? In 2016. There so everyone knows about the Dow and I think you know, the Dow is fascinating in the Dow was kind of fascinating for a product protocol politics perspective, because there were all of these debates about how, you know, is immutability, like in an unlimited strong role that use Like to in every situation like, is it like a kind of Fiat you Stacy up here yet Windows sort of thing? That's a Latin phrase for like, Let there be justice and let the world perish? or might there be like extreme case limits on it? And, you know, there's a debate of like, would that end up setting a precedent? And I think the answer to that actually ended up being much less than people thought, right? Because a year later, there was another dispute, the eip 999, where the parody team tried to unstuck their coins and like, make a hard fork to do that. And the community just said, like Halibel, like we get no, we actually explicitly don't want to create a precedent and or we explicitly don't want the Dow to become a precedent. And so they kind of intentionally what's in the other direction. And since that there haven't been really any attempts to violate immutability to anything like what happens with Fidel, but

Balaji Srinivasan 15:56

I want to just ask one question on EIP 1559. Because I do remember this, and I'm actually looking at it again. And it looks like the big difference is, before it was an auction, which was in Well, first, a first price auction, as opposed to like a second price Vickrey auction is suboptimal. But the second thing is it just introduces noise and unpredictability and everything because you have a fee market and so on. And it looks instead that you went to a protocol set base fee, so that if you pay that base fee, you're guaranteed of being included, and so the unpredictability is somewhat decreased. And I'm looking also at block times, and they've also flattened out and the variance has gone down since the merger was a little bit more variance before the merge. Is that correct? Or am I off base on that? Yep. Yep. No,

Vitalik Buterin 16:36

both of both are very correct. Okay. Great. And actually,

Balaji Srinivasan 16:39

it's interesting, because I, you know, I known about EIP, 5059. I hadn't like tracked it closely. But it's interesting to see that that combination of things, is what has been responsible for kind of faster, and more predictable. transaction fee inclusion on the Dow. Do you want my perspective on the outside? Just watching the whole Dell Dow thing? I didn't hear that? Go ahead? Yeah. So first is during the Dow, you were the and for those folks who don't know, because this is now ancient history and crypto terms. It's seven years ago, it's 2016. When Aetherium was I think Bitcoin was starting to hit congestion problems. And a bunch of folks were checking out Aetherium in early 2016. And it had gotten up to a billion in valuation. I think there abouts in early 2016. There was this project, the Dow that was trying to do effectively, an on chain VC fund. But it was trying to be autonomous and so on even the name Dow is probably arguably antiquated, it would be really, it's not really an autonomous community. And it's, we're not autonomous organization. And so from the outside, what happened was, I think was like 10%, or something of all eath was in the Dow, was that right? It was

Vitalik Buterin 17:45

actually it was 11 and a half million E, which is a bit under 10% of the East that exists today. But it was I think, 14% of that, if that existed back then.

Balaji Srinivasan 17:55

Okay, so it's like a big chunk. And so what happened was this, you know, Dow attacker read the code and figured out some way to sort of drain the the DAO contract, but there was this thing where there were like, there's like a white hat approach, which could keep going back and forth and keep, like, undoing his thing. And that was just going back and forth. And I think there's like a 30 day clock or something before you before was like a final transaction or something like that. Was that is that right? And then you're gonna get into Patch before that three day clock went up to stop those funds from being moved somewhere. Am I remembering that correctly? Yep. ballpark. And so thing about that was, I remember from the outside Washington, I was like, you know, now I have sympathy for Bernanke. Because there was a systemic risk to the whole system. And this sort of intervention could save the system potentially, but would it create moral hazard and so on and so forth? or what have you and I, I just understood, in the same way that when you start a company, you're CEO, and you understand what it is to be CEO or you, you teach a class, your professor, you understand what it is, and you're like, Okay, now, I know what the smartest kids were on the other side, right? Being actually in the position of essentially central banker, you're like, Okay, I now understand why, you know, they're there. Did you have those thoughts at that time? Or is that something you're thinking about?

Vitalik Buterin 19:11

I don't think I use the Bernanke as an analogy. But like, I guess just because the those kinds of like monetary and finance issues are probably a little bit less of a primary category for me, and then they are for a lot of the other earlier crypto people, but definitely similar thoughts.

Balaji Srinivasan 19:32

It's, it's interesting, because I mean, ultimately, the justification is that Aetherium has succeeded, and that there hasn't been despite significant pressure, other sort of reversal things put into a theory. Like of course, there was that unfortunate hack of the parity wallet, for example, and then they were trying to get an EIP in there, I believe to like, move the funds wherever but so far, while that's very unfortunate, I don't think there has been another thing that's like that It is like allowing for community reversals of transactions. So that means that we have stable governance. I guess we're out of homestead fully. And you know, was it formally in Homestead around the time? The Dow

Vitalik Buterin 20:10

is yes. So homestead happens in March 2016. And the Dow got hacked on. I believe it was June 17 2016. And yeah,

Balaji Srinivasan 20:20

so it's still basically it's still basically beta. So that's, I think, a fairly good justification for beta versus prod as to you know, when edits. So now now we're post merge. I mean, and I was, I was just crossing my fingers, actually,

Vitalik Buterin 20:33

before we get to post of our agenda. Yeah, I kind of wanted to talk a bit about the other thing that happened in 2016. And just because I think it's, it's equally fascinating, but like 10 times fewer people will have heard of it. So okay, so the time is 2016, September 19. And in the background of this time, was that the Dow had gotten hacked once ago, and the Dow hardfork that had, like, done that kind of surgical intervention to the chains. They're just like, delete the Dow coins and move them into this other place of how had happened about two months ago. The whole situation where the Etherium classic community rebelled and made their own version of Aetherium had that refuse to implement that change had happened, and the entire kind of market drama where, you know, you had Barry Silbert, polpo push trying to push Aetherium classic, really hard a whole bunch of characters kind of buying in early trying to get it to flip Aetherium when you had a whole bunch of exchanges supporting yet Oh, just a whole bunch of people really like watching and seeing, you know which of these two chains would win because this was the first time there was any kind of like, literal crypto civil war between the two chains. So that had all happened. And the BTC price had hit a maximum of 0.4 eath. And then it had dropped again. And it had dropped to think a bit below 0.1 at the time. And it just felt like maybe just maybe I don't know, the theorem, community was finally on its way out from the being in the kind of Dow arc of the plot of the story. But then, at 5:15am, Shanghai time on September the 19th, when I was staying in the hotel, where the DEFCON two was literally about to start in three hours into 45 minutes, I got this military style wake up call, like just woken up by a call, I picked it up and someone from the theorem foundation and said, hey, you know, we have a situation he needs to come down now. So I pretty much stood up, grabbed my computer, put my computer in my computer bag, walked out of the hotel room, elevator down, walked over to this room where I was supposed to go. And I saw about, you know, 20 Aetherium developers gather the room. And at that time, they were still trying to figure out what was going on. Basically, what had actually happened was that someone discovered a bug, not in the Etherium protocol, but in Aetherium implementations that would cause it to process a particular type of transaction much more slowly, like basically, instead of taking like n steps to process a transaction that had an computational steps in it in this one particular case, it would take n squared steps, right, basically, because it was like each single step of the EVM would add one extra thing that the that the node would need to worry about and the n so the first thing you would need in the first step would need to worry about one thing and the second step, you need to worry about two things and three steps. Are you worried about three things right? And so when quadratic like this kind of bug could totally have actually, a quadratic bug actually did happen in Bitcoin? Right? So one of the justifications for DOD is what are the secondary justifications for not increasing the Bitcoin block size? Back when the Bitcoin block size debate had happened? This was one of these like technical blog things that again, all of a sudden no one knows about, right is the quadratic attack issue with hashing? So the quadratic attack issue with hashing in Bitcoin basically, yeah, it is. I think it's, I think it still is, but the only for oldstyle transactions segregated witness fix that, but basically, that if you have a transaction that has n inputs, then to compute the hash that gets used for the signature in each of the inputs, you have to hash the whole transaction, read it. So if you have n inputs, that each of those n inputs, theta n different signatures, you have to create n different hashes in each of those hashes contain, you know, at least kind of Like the structure of all of those inputs, and so each of those hashes are besides n, m ID. So you would have n hashes of size and in order to sign a size N transaction, so you need to O of n squared hashing, right? And because Bitcoin is just like because it was just hashing and not EVM execution as because Bitcoin just has a lower absolute limits to its capacity, it ended up being I think, okay, for Bitcoin as it as like a worst case block at that time, I think ended up taking like 30 seconds to process. But if the block size had increased, that would go up quite dramatically, right. So if the block size has gone up from one to 10, then the the worst case block time would have gone up from 30 seconds to 3000 seconds, which is almost an hour, right. So. So it's not a Turing completeness issue, right? It's a quadratic, like literally Bitcoin also had a quadratic execution issue, and that ended up influencing, at least to a slight extent, I think it ultimately happened for other reasons, the Yes, wall block versus big block discussion. In the Aetherium. case, it was, it had to do with the implementation and way the way that the system handled recursive calls, or just like log towers have calls in general. So like a calls B, call C, or it could be a calls a calls a, and how it creates like, basically, when you created a call to VM ended up copying everything, and the the amount of stuff that got copied, kept increasing with every call. So there's actually a whole bunch of these different quadratic execution attacks. So the attacker discovered the first one in an early morning, Chinatown time of September the 19th. I think the attacker might have even intended to kind of create a nasty surprise and to basically make the theory of chain break on the morning of DEF CON. But we ended up working really hard and really intensively for those few hours we had, by about 6:30am, I think, isolated the bug by like maybe seven 8am or so created the patches that would fix the bug. And that's pretty

Balaji Srinivasan 27:12

legit. That's hard to do for a subtle bug that's actually quite difficult. Go ahead.

Vitalik Buterin 27:17

It absolutely was. Yeah, then by like 8:30am, we had, I think, published to the patches, and people were starting to download them and add a nine to 20 5am, it's 25 minutes after the intended start of the conference. We were all kind of confident enough that everything was stable, people were running the new nodes, everything was working as expected, and the network was running fine again. And the conference basically started pretty much as it had been planned to start. Right. So we had turned a in a basically an attempt at creating a PR disaster for us, I think, you know, into something that really showed the unity and the just the dedication and the competence of the gift developer team. Right. I think it's one of the other reasons why I kind of wanted to tell the story, which by the way, it's not done yet. It was that, like the core Devs really are these kind of unsung heroes of the ecosystem. And they tend to be more introverted, and they tend to be less podcast D then you know, myself for just that, but they've just done a lot to make Ethereum Ethereum clients faster. And to really enable like, let's say, the 7x capacity increase that the chain has had since since it launched right, but it took a long time to get there. Right. So that was the first attack. And then when DEFCON finished on the evening of the third day, the attacker discovered a second attack. A few days later, the attacker discovered a third attack, there was a fourth attack. There was literally this month long period we are we were literally in constant cyber war between the guest and parody development teams a gift and parody were the two Aetherium clients that were functional at the time. And there would be often periods where either death works in a parody to not work for parity works and get the network and a lot of users ended up switching between the two clients so really frequently. And the two teams were both just like frantically putting out fires that this attacker was starting and he's hacker was figuring out every single bug in each of these ethereal clients very smart guy or the attacker was Yeah, his really impressive right and then finally, a month later, he discovered a bug which was not in the implementation, but in the protocol itself. Right. This was a bug with the self destruct opcode that basically allowed it to be used to do some real would basically make the Ethereum stage much much bigger using a very tiny amounts of gas. And that and that bug had been left alone like it would have been fatal. Within some number of weeks, right, because basically the Aetherium state size was just blowing up. And once the theory of state size bloats up too much, then there would just be no way to do any kind of caching. And that would make a theory of even slower in other ways. There, we had to do another hard fork and other emergency unexpected hard fork to increase the gas costs of a couple of operations to just make those attacks just not viable in the future. Do you remember that? And do you remember the second hard fork that added a feature to clean up all of the junk state objects that the attacker added into the theory on chain, and I

Balaji Srinivasan 30:38

remember basically, that at the end of 2016, in December, like a theorems price had dropped quite a bit over the course year has been like a long slide down, people were demoralized after the Dow and the multiple hard forks. And people thought that the This is not you. But folks, you know, like who are in the broader crypto community are thought, okay, smart contracts are just too hard and dish add too much complexity, too much attack surface, it's going to be a while before this thing works, and, you know, much, much power to them, but it'll probably take a few years. Then three months, four months later, you have this gigantic rise. And it's like, you know, we should play that wrap the you know, where the guy is like, on his back. And he kind of gets up like that, you know, that meme I'm talking about? We'll put this in the show. And it's Go ahead.

Vitalik Buterin 31:23

Yeah. But you know, the the the trouble was, that was not quite over, as you said until December, right. At the end of November, there was a situation where like, I had to send a bunch of transactions to me basically wrote a Python script to send 1000s of transactions to clear out this junk that the attacker had added into the, into the chain. And I'm like they, and one of their transactions triggered another bug that had accidentally been introduced, that actually for the first and only time like really act proper would actually had a consensus failure, like same thing as what happened in Bitcoin in 2013. He kind of split the chain in half. Like I was the one who sent a, the the transaction that like basically broke Aetherium in that way for the first and only time, which was one of these kind of fun stories again, but yeah, after all that the price had been dropping, everyone was definitely demoralized, and then 2017 happened and the crazy bubbles of 2017 happened, then 2018 happened. And that's when the 90 and a happens and people went to theory Mortimore realized, again, the end of 2018 was the birth of uniswap, which I think was a good kind of segue into the defy era. And then in 2020, we have the summer of Defy. And then there was started to be NF Ts. And if these are probably the one thing that I did not project, by the way, right, like, if you look at the list of applications that were in the theory on white paper, and you'll look at the applications that are popular today, the big thing that's in the second lesson, that's not the first is that NF Ts, but otherwise, you know, their prediction markets are defi. There are a lot of things, right. And then finally, you know, this last year 2022, the merge happened. And that was serenity. Right. Serenity was always intended to be proof of stake. And later on, we added on sharding, the basically some kind of scaling solution to that, and no proof of steak has finally happened. Congratulations

Balaji Srinivasan 33:21

on so can I say why I was concerned about that? I was I was just shocked. I obviously, you know, rooting for the success and everything. But I was like, you know, a full switchover of the of course consensus had been tested for about a year or there abouts. But you're moving the whole ecosystem over. And it was a everything that isn't tested is usually where bugs arise. And so the fact that that was as smooth as it was, and basically a non event, and that there hasn't been I mean, maybe there's something that eventually but it certainly nothing at the time, nothing in that few weeks afterwards was really impressive. It was is one of those back end changes where the whole point is that it's a non event, but a very, very, very difficult to pull off. So congratulations on that. You probably have some thoughts on that whole transition over

Vitalik Buterin 34:13

Yeah, this transition took a lot of work right. Like I think it actually surprised all of us that the transition wasn't an event to the extent that it was right because every single test run that we had attempted had at least one tiny thing break in it like never anything bad enough that like it would have been a big problem had it happened on the on the live network, but it was always like some combination of client settings ended up not surviving and like 10% of the nodes dropped off. One time 35% of the nodes dropped off and that actually caused finalization, it's away by an hour. There's just like always a couple of one or two tiny things that would break and then when the real merge finally happens, you know, everything just went extremely smoothly. Right. And I think that again, you know, it's just, it's a testament So the quality of, you know, the all of the theory on client development teams that worked hard for multiple years to write the software that would make the transition to proof of stake possible, right, and basically, pull off something that I think some people were describing as changing the engine of a jet while the jet was still flying. Right. It's pretty, really complicated operation. And a lot of people I think, definitely did believe that it couldn't be done. But I'm, you know, eventually it was done. So, so now that the Birch is done, and I think the main thing that's the left for the for in the original kind of roadmap is sharding. Right. And sharding is something that we talked about for a long time. But I think the ideas around how sharding would be able that could have changed, it's gotten more ambitious, right? Like, early versions of sharding, were willing to trust honest majority assumptions. And basically, we're willing to trust the idea that, like, let's assume 50% of the nodes are going to be good people, which is, is like a lot less ambitious than like, say, the Bitcoin spirit, right? Which basically says that, you know, hey, even if 90% of biters are pushing the wrong chain, you as an operator of a node should be able to reject it, right? Like bitcoin is not a minor oligarchy. It's like a constitutional system where it's supposed to only accept things that are valid. And the original vision for Aetherium sharding had been like, Okay, fine. No, obviously, we have to trust the 50% are good, to be honest. But then we started being more ambitious and saying, like, Oh, can we actually make a system that supported many more transactions that a single node could verify. But that was still verifiable in the sense that the system would still kind of force itself to follow the rules, and where users could verify that rules are being followed? Even if a great majority of the consensus participants were trying to subvert that objective, right? There's a lot of really interesting research and what have you be? And I think fate.

Balaji Srinivasan 37:04

So what percentage would be trying to subvert in that case? So not even? Like, what's the honest, it's not the honest majority anymore? It's an honest fraction of what?

Vitalik Buterin 37:12

Yeah, right. So we had like an honest minority assumption in some of our protocols that basically required like the user and maybe a few 100 other other number, not even a fixed

Balaji Srinivasan 37:23

percentage,

Vitalik Buterin 37:23

basically, no, org member. Right,

Balaji Srinivasan 37:27

exact. That's really interesting. Okay, I actually, I should look at that. Go ahead.

Vitalik Buterin 37:31

Yeah. But it's one of the fateful decisions that we made around 2020. This is, I think, another one of these trivia that's, that is well known within Aetherium circles, but probably hasn't really kind of filtered out to the, you know, the wider narrative meme, meme blocks is the roll up centric roadmap, right? This is basically this idea that, instead of Aetherium, itself, trying to kind of boil the ocean with scaling and insert itself into this complicated hyper scalable system, we would make some changes to Aetherium, that would make it be more friendly to people making layer two scaling solutions, right. So layer two scaling solutions are like protocols that try to do the same thing that Aetherium does. But that plugs into Aetherium for security without doing every single thing on top of Aetherium. Right, so most of the work would happen off chain, but just enough work would or could be done on chain, that if something happens, the Ethereum system would kind of keep the rest of the system honest. Right? So one of the defining analogies for layer two is in general, like the Supreme Court, right? Like, the concept of exactly the Supreme Court is, what things are going to ultimately rests on. But you don't literally have the Supreme Court hearing every single case, right? You have a whole bunch of these appeals courts and like a whole bunch of lower courts, and then you just have the fact that the threat of a court case being possible can drive on this behavior without most of that behavior effort needing to get into a court system, right? That kind of analogy, like it drives the Lightning Network, which is basically the only kind of layer two that's possible on Bitcoin. And it also drives a to a state channels, which are like the Lightning Network level that's on Aetherium, as well as roll ups, which are this more complicated thing that was only discovered later on a theory that are actually powerful enough that you can do smart contracts on top of that, right. So Lightning Network has state channels support transactions, and they support like very limited kinds of contracting, the roll ups the support like kind of full on EVM that are that kind of get indirectly verified inside the EVM. So layer twos have kind of been the theory of scaling direction since then. And since then, there have been these discussions around data availability, right basically If the practical capacity limit to a layer two is how much data, the Etherium chain chain itself can store, right, because one of the trade offs of roll ups is that they do require the theory of change a store a few bytes of data for every single transaction, like a lot less than using a theorem directly, but still a settlement of data. And that's where stuff like data availability sampling and EIP 4844, which is looking I'm scheduled to come early to bed juncture, no, maybe maybe later next year is is all about exactly. So 3 million will predict sharding is like a bit of a plan, right? Because So first there's sharding. And then there was dank sharding, which is this version of sharding, which was created by the theory. Mr. Researcher, name's Dan grad. And then there is proto dank sharding, which is called that because it's a stepping stone to full day charting. It doesn't technically shard data yet, though it does shard history and computation. But it was also created by another theory of researchers famous proto lambda, right. So, yeah, so first game sharding happening next year. And the other thing that, like I've talked to the roll up teams, and they all want to do next year is they want to start taking off training wheels, right. So the roll ups and layer twos that exist on Ethereum today, they basically all have what I call training wheels, like some kind of backdoor that lets developers come in and like say, stop and change the protocol, if they see that some kind of bug has happened. And like training wheels are obviously a, you know, an affront to the moral idea of trustful status, and all of those things, and nobody wants training wheels to be the status quo long term. But basically, they're like a compromise saying, Okay, people up, like, we want to hit the ground running, and we want applications to be able to start, you know, start deploying all your tours before we can be fully confident the thing that the thing has no bugs. And so we're going to have a backdoor for a while. And then once things are stabilized, and once we're confident enough in the code, then we can finally start screwing the backdoors out, right. And so next year, I Yeah, some of the projects are going actually, I think might be actually taking their backdoors out, some of them are going to be kind of limiting their backdoors. And going through a hybrid, right where a backdoor exists, but there's like, it's not like one person that manages it. And it's not even a multi SIG with a 51% threshold. It's like a multi SIG with an 80% threshold or a 75% threshold with members that are like highly distributed bids between a bunch of different organizations. And so it's like a halfway house between, you know, being a kind of trusted sidechain, the, in the style of, you know, something like liquid or the style of something like polygon today versus being a fully decentralized system, right? Where it's like, you know, the humans have half the votes, and the code has half the votes, that's probably one sort of mathematically accurate way to think about it. And so the Senate in the house, if you want to provide the code, yeah, exactly. It will, like if the, if even a majority of the humans are dishonest, as long as like, unless you have, like, almost all the humans are dishonest. If it's just the majority, then you know, the code still kind of carries the day, right. But so the idea would be that if there is some political controversy, that then you know, you're not going to, you could see yourself getting 51% of the humans, but, you know, you're not gonna get like 80% of the humans. Whereas if it's a bog, then like, that's something where it's easy to get, like 100% Even agreement that the bucket needs to be fixed, you know, not total kind of, you know, take the training wheels off and weave into the code, but like a very large step toward that. And actually, one of the teams I talked to actually is more ambitious than even that one of the their approach is to basically say, we're going to have a system where we only turn on the humans, when someone submits a proof that the code disagrees with itself. So if the code can create a valid proof for something that's valid, and an ad for something that's invalid,

Balaji Srinivasan 44:23

then it's inconsistent and some sushi humans like,

Vitalik Buterin 44:26

exactly if you make a proof of inconsistency, right, which is much easier than a proof of incorrect this, like, basically just show that it provides two different answers to the same question, right, like, you know, if, like, if I asked Chad gbta, you know, who was the president of the United States at 1850 and ask it 10 times, and one of those times it gives a different answer. Like, that's, that by itself is proof that that it's not, you know, a perfect system this like, even to someone who knows nothing about US history, right? It's like that kind of idea. So, but this is all like I I think the details of that are still being figured out. But like, the thing that everyone is dead set on is that 2023 is going to be the year when, you know, roll ups really come to maturity, we have 44 Four that gives the ropes more space, and we have the training wheels on roll ups, like really get weakened a lot. And like, basically yet, we go most of the way to these, you know, a theory of scale, like the end of this kind of fully trustless. And so not fully trust was most of the way to being forged by sort of symbols, which mostly trustless system that kind of gets us to, like most of you know, what we would call sharding being finished. And then after that there is roll ups going fully trustless and then there's full tank sharding. And I think those things are going to take a couple of years longer. But like, for me personally, I think after the changes in next year, like that would be the point the first points where I would feel comfortable saying, even if nothing else happens past that point, I would be happy with Aetherium. Right? And that to me is like the like my Yeah, like that's the personal milestone for me, right? Like the, you know, equivalent of like, when can we call a theory I'm done if we really have to call it done? Well, the end of 2023, right. So merge done, withdrawals done. So being able to withdraw from the proof of stake system, you know, basic scaling done. A, you know, a lot of basic improvements done enough cryptography added to a theory that we can support privacy solutions. So layer on top. And you know, if we need to, we could stop there. But as you know, we're not going to stop there. Right. And we wants to kind of like, go ahead and try to actually make sure that the theory ecosystem gets support 500 million users before, you know, the bowl were 500 billion users are knocking on the door, actually, yeah. You know, ends up happening. That's great. So that's, you know, that yeah, so that's that that's probably the next part of the future, you know, the transition from Ethier Iam as this kind of fear, very theoretical ecosystem that's still discovering and finding itself to Ethereum as an ecosystem that actually is, you know, tries to be useful and usable and, you know, secure and, and all those things and like, actually provide value for hundreds of billions of people.

Balaji Srinivasan 47:21

Here's a question which you can skip if you want. But basically, in terms of Aetherium being done, so you mentioned, you know, the trust lessness. You know, some folks have talked about like the OFAC compliant blocks due to, you know, the whole MeV issue. Others have said, Oh, when or when are we going to be able to withdraw Aetherium from staking? Do you have any thoughts on? And then the third is basically, you know, I think Zuko has mentioned that it'll be hard to have privacy completely bolted onto Aetherium. If it's not there at the base layer. And I think I might agree with that third one. But maybe do you have any thoughts on those three? Or do you not want to address those or? Go ahead?

Vitalik Buterin 48:00

I think they're definitely important issues. So withdrawals are the easiest one like it's uncontroversial. Everyone agrees that it's going in the next hard fork, which is, I think expect to happen early next year. All the like privacy and censorship resistant stuff IV N, it's, there's definitely like, a big bundle of problems that I think the etheric medical system is facing at the same time that are important, right. So it's, like basically, one of the things that's happened over the last year or a few years is the rise of MeV. And MRV. Like, one simple way to think about it is it's like the fact that there exist strategies for miners or block proposers or whoever creates the block to make even more money, where those strategies are complex and require you to have like complex algorithms to be able to execute on them, right. So it's basically the fact that like, being a really dumb automaton is no longer the optimal strategy for a block builder. The challenge is like, how do we prevent that fact from being a source of centralization? Where, you know, let's say staking ends up centralizing into like two or three pools that get become better than everyone else at actually implementing these algorithms. Right. So that discussion started happening in around 2019. and flash floods came out around that time and then at the same at the same time, you know, they're definitely yeah, we're mining pools that were independently thinking about how like they would try to extract this extra NPV revenue as pools. So that that was one of the things that happens. And the way that we ended up solving that problem is basically this concept of proposer builder separation right where, instead of the, whoever creates the blocks, the proposer, as being able to, or needing to figure out the contents of the block themselves, they can auction that off to third parties, right? And they can basically run an auction. And they can accept bids that say, Hey, here's an idea I have for what you included a block. And if you include this, or that, you know, I'll let you let you include this and collect the fees. And if you'll let me do that, I'll pay you 0.018 eath. And someone else might say, Oh, here's another block that I have. And if you include this, I'll pay you 0.022 eath. Read. And there's like a lot of complexities around how to make that kind of market fair for all participants and like make it decentralized, that resistance to cheating and attacks on all on all sides without the resistance, or without the anti cheating itself turning into a centralization vector. But one of the things that happened as a result of this is basically that yet we ended up outsource, like block proposers ended up outsourcing too much, right? Like, you don't actually need to outsource the entire process of block creation, if all you want to do is to capture MeV revenue. And the problem with outsourcing the entire process of block creation is that we've moved block creation from this set of for distributed actors to this set of more concentrated actors, and more concentrated actors are inevitably going to be much more legally conservative and even more likely to over comply, in terms of what kinds of transactions they are accepting, right, like, basically, we now have a bell which 70% of block of the builders that are creating, that are actually responsible for constructing blocks that are going into the chain that are refusing to include certain kinds of transactions, like I've actually, you know, like even like talks to legal counsel that think that they're going too far, and that they're even, you know, blocking transit transactions to an extent, to an extent where they don't need to write, but no, the fact that they are centralized actors, the fact that, you know, even though lots of them are, you know, based in the US the fact that they are actors that are connected through other actors and all these things, right, like, it ends up kind of pushing things over to the, you know, to this conservatives side that, like, you know, goes beyond what a lot of legal counsel even, have you even told me as legally necessary. And so the good news is that nothing has been censored in theory, right? So the good news is that even the transactions of the type that people are being most conservative about trying to censor those transactions, still, on average, getting booted into the chain after about three blocks, right? So actually, post EIP one, like basically EMP 1559, it did more to make those just to make those transactions faster to include, then the whole IBV situation, is there some slower to or

Balaji Srinivasan 53:03

is there some Canary for Ethereum transaction censorship? Like could you there's probably something there, but maybe I just don't know the site that's just looking at the mempool or, and seeing that there aren't blocks or there aren't transactions with valid high amounts of fees that aren't making it in?

Vitalik Buterin 53:20

So the yes, the canonical site for this is MeV. Watch dot info. That's actually the site where I pulled the 70.

Balaji Srinivasan 53:28

Yeah, I do remember this, the OFAC of blind blocks the top right, but the censorship offenders says censored. So it does actually claim that some of them are censoring

Vitalik Buterin 53:41

these entities run censoring and if you release other validators, yeah, but it's improved.

Balaji Srinivasan 53:45

Like, I have not seen this graph at the bottom layer, right? The posts emerge daily. So, so it looks like green is increasing, which is interesting and a little bit counterintuitive. Is it your view that that's going to continue to increase or that green is large enough that that's the free Etherium for a while, and it just needs to be large enough for how do you think about that? I think

Vitalik Buterin 54:03

we're expecting green to slowly increase over time, right? So by the way, just like my view on like, Is it our Are we in a state that's good enough or not, is that one of the nice things about blocks via a transaction censorship in blocks is that it's what I call a one event trust model, right? So even if you know you have 99% of builders, or block proposers connecting to builders that are censoring, that just means that someone has to wait 100 blocks instead of one and that's still like, kind of fine for them, right? Like basically, if you if you want stuff to be censored, you'd have to, like reach pretty much everyone, but at the same time we had or you'd have to lean on pretty much everyone and get them to stop including uncensored blocks, but at the same time, like you know, given how things like you know, normal cascades work and just all those social phenomena like we're definitely not had in the Santa belt this right so, like just to give something concrete examples of Are the theory of ecosystem being not head in the sand? One is that a multiple view entity is announced in, I think, is in December in January, that they started operating? No, I did not realize right and Yeah, and like a lot of them are even, you know, entities that are, you know, outside the US serving stakers that are outside the US and all that, and all that. But, you know, they did watch, and some people have switched over to them. But it's, it's like one of those slow processes because I mean, no staking is like a multi billion dollar thing. And everyone who participates wants to be just careful about moving over infrastructure and all that. So that was the first thing. The second thing is that flush flush toilets team added a setting to NBv boost that allows you to say, I'm only outsourcing block production, in those cases when outsourcing the block production is significantly profitable, right. So if the revenue is like less than point 05, or some threshold that you can still make blocks yourself, if it's over point oh five, then you can go and outsource. And it turns out that if everyone says point 05 as their fee setting, then you would have a world where 40% of blocks would not be outsourced. But every single validator would gain like something like 96% of the revenue gains that they would gain from outsourcing all of the time. Right. So this is like one of the examples of what I mean when I say that people are outsourcing too much. And that's a setting that people are starting to use. And then in the longer term, I think one thing that we want to do is this concept of partial block auctions where in basically, instead of it's like when you outsource you set conditions. So you say for example, I have seen these one of these 23 transactions. And so what you create a block, I want you to include all 23 of these transactions, or I'm gonna like you choose the first flush transactions in the block, but I'm going to choose the other ones. And if you do this, then like, you're still going to be able to get like 99% of the revenue from full outsourcing. But because 99% of the revenue from full outsourcing really comes in like the first five or 10 transactions, but at the same time, you know, like, if everyone does that, then like for me censorship perspective, that's great chain is going to be fine. Right. So that's the base. That's the base chain. I think the other thing that I think is worth talking about is like, the application layer perspective on privacy, which I think is important, like privacy is super important. But I do think that there's like, there definitely are these kind of extra legal challenges that have to be that have to be navigated, right, like basically. So you know, it's real cash exists, and tornado cash still exists, right. But like, even if tornado caches is still completely, uh, you know, usable and unsafe and uncensored from an ethereal political perspective, that does still, in practice, like, like, the legal actions have in practice a little bit of its usefulness right. And they've limited its usefulness for two reasons. One of those reasons is that it's become extremely illegal or risky to use for US persons. But the other reason is that, if you receive a tornado cash coin, you're going to get a lot more trouble from like exchanges and the entire rest of the kind of crypto Fiat bridge ecosystem that needs to comply, right? Like one approach to trying to kind of deal with that is to basically say, you know, we're going to make privacy a default for everyone, and we're just going to make it, you know, impossible for anyone to distinguish themselves from anyone else that wants to try really. So

Balaji Srinivasan 58:56

that's actually where I was gonna go is basically that, and maybe we can do this and then we'll move on to other stuff besides crypto stuff. But it seems to me like something where, you know, whether it's commit reveal or zero knowledge commit, there's like a bunch of different dynamics, but were minor miners can't see what transactions are including, nobody can see what transactions are including it takes away MeV. That's a justification for on a completely different grounds. And you're sort of privacy baked at the base layer where you intentionally make miners dumb pipes, in a sense, right? They're blinded to it. And there's probably some clever, cryptic, you know, cryptographic tricks you could pull to make that feasible. And then you also support zero knowledge for the transactions themselves, not just the, you know, the mining process. And so something like that, where the system cannot, quote, discriminate against transactions because it cannot tell the difference between them. Seems like where you ultimately want to go or, you know, tell me tell me your thoughts on that.

Vitalik Buterin 59:56

Yes, I think at protocol level moving in that direction will definitely be here. A really interesting, but I do it on male too. Yeah, but But it'd be in like application level, I think it's more challenging, right. And application level, I think the like, I mean, I know this is a very controversial topic. And, you know, people have, you know, all kinds of opinions on this. And I think, like, my view on this is probably a little bit more willing to compromise than some of the purists which is that like, even though, like, I think that the Etherium chain should like, it should absolutely have, like layer one censorship resistance as the hill that dies on and, you know, if it has a choice between, you know, censoring on L one, and, you know, even, like, a 50% chance of getting killed or like you should even take a 50% that take the 50% chance of getting killed, because what's creative, what's credible neutrality is of the chain is broken, then what's the point? But on the application layer, like, like, first of all, you know, I think, you know, there are legitimate concerns, right? So like to, you know, it is true that a very large portion of the inputs to 3d to cash ended up being these large scale sci fi hackers, you know, it is true that there are a lot of large scale criminals who are using these using these systems, and who would use these kinds of systems if they ended up even big, kind of becoming much larger unchanged, which

Balaji Srinivasan 1:01:35

one of the largest criminals is maybe supposedly the North Korean government? So would you say government was one of the largest criminals? Just keep going?

Vitalik Buterin 1:01:44

Very possible, you know, and but it's, and I think, just also from a philosophical perspective, right. Like, like, if I had to choose, you know, what would the rights kind of place to be in terms of financial privacy? Like there's this? So there's this slogan that, you know, even I think the cypherpunks originated, like 20 or 30 years ago, which is, like, what we want is transparency for the powerful and privacy for the weak, right? Yeah, like surveillance, but fumble? Yes. Right. Exactly. And I think like, even outside of law enforcement use cases, like, you know, people appreciate the fact that they can do kind of open source on chain analytics and like, track the crazy billion dollar stuff that's been happening inside of FTX. Or, you know, what some of the other large exchanges are doing or whatever, right. So I think, what would be, what would be nice, you know, is a world where there was something approaching complete anonymity for small transactions, and, you know, very high privacy for medium transactions, but where larger transactions and especially transactions that, you know, huge amounts of people don't, don't even you know, wants to make an attempt to mix themselves bandwidth, don't necessarily have that same level of privacy. And the interesting thing, right, and the interesting thing about zero knowledge cryptography is that it actually lets us do this, right. So here's like, the simple way to do this. So imagine a version of tornado cash, we are when you withdraw from 22 cash, you make one proof, which is proving that, you know, you have a valid coin that's deposited into the system, but you haven't done yet, and you haven't withdrawn yet, which is the thing that's been proven now, but you also add in another proof, which proves that you are not one of, let's say, 23 addresses that, like chainalysis, or some collection of organizations decided, you know, our large scale, possibly North Korean hackers, right? So like us users decide to say, Hey, I'm willing to mix myself in with the randos. But I'm not willing to mix myself in with large scale defi hackers, and contribute to large scale defi, hackers, anonymity sets, right? And if let's say, yeah, 95% of users take this option, then the anonymity set of the worst multiplayer hacker reduces by a factor of 20. Right. And so that makes like privacy for large scale. defi hacker is much smaller without affecting privacy for everyone else. And, like, the really nice thing about it is that it's incentive compatible, right? Because, like you as a user, you know, you wants to, I think proof that you're not a large guilty if I hack or both, because, you know, you as a user, you know, think morally don't want to support that kind of behavior. And because you as a user, don't want to get this mistaken for a possible defi hacker and, you know, get a higher suspicion score or whatever because of this, right?

Balaji Srinivasan 1:04:42

I've ever seen the ending of The Thomas Crown Affair. It's the it's the scene where they all are wearing hats, and they're indistinguishable from each other. And I think what I just, you know, Zuko has talked about this as well, but it's not a private transaction. It's a privacy set, right? That pool in which you're gonna so I think, you know, your comment there just made me realize that privacy is actually a collective good that a group can choose to give a member or not. It's not that it's not just a technical thing it is that because obviously, there's cryptographic kinds of privacy where you can't invert a function, you know, chairman who wrote some, like, you know, let's say you hash a document, you can't just invert, that doesn't matter how many people around it. But a collective privacy does exist as well, for some use cases. And thinking about that as a tool that you can dial up or dial down based on consent is interesting. Just Just interesting point I had known about in the context of Z cash, I just hadn't thought about it in the context of something that the members could selectively do, actually, you know, in ring signatures, that's more common where people are choosing who's part of the ring, right. But I just hadn't seen it in just this particular context, which is interesting. So do you want to switch gears for a second? We've talked about crypto for about an hour, we'll spend one more hour on other stuff. Alright. So second, big question. First big question was basically what's going on with Aetherium? Which we know about the history? What do we know about you know, what happened with the merge and the all the stuff in 2016? And, you know, how does a theorem actually become a final product? How do you get past? You know, the issues with OFAC compliant miners? And how do we increase freedom there? How do we get past the withdrawal issue? There's a pull request, or you know, there's, there's gonna be a hard fork on that early or next year. And then how do we add more privacy? We talked about all that stuff. Okay. So Now switching gears. Here's a big question. Ready? All right. So metallic, you have, you know, as I've said many times, we've started new companies, we started new communities, and you yourself have started a very important new currency. So if you could go one more level, what new country would you start? And why? Hmm.

Vitalik Buterin 1:06:59

So I think it's important to approach that question, not from the, the kind of LARPing point of view, like I wants to have a country because countries are cool, right? It's like this, like, it's actually important, I think, to kind of dig down into, you know, what is a country break it into its constituent parts? And which of those constituent parts do you need? Do? What do we need? And what ticularly know which of those parts? And what combination are valuable for people in the context of, you know, being just out of the grand old here of 2022? Right?

Balaji Srinivasan 1:07:38

I mean, the serious version is, what is the single biggest problem the world that you'd address with that new country? Or would you, for example, disagree that there'd be a single biggest problem as per our discussion the other day? Maybe you can go into that.

Vitalik Buterin 1:07:53

Hmm. Right. So I think one of the ways to kind of break down a country, right, is that there is the social network, and there are the legal and jurisdictional properties, right. And I mean, across possibly a third one is like physical location properties, right. So like, for example, if an even if, you know, you decide that the legal environment, environments and even cultural environments of like, you know, Dubai, for example, is, is perfect for you and like, which, like I personally, you know, don't think of it. I think there's a lot of a lot of upsides, definitely, but also a lot of downsides. So there, there are two, right, but like, the, the really big one is climate right now, its climate is not their fault. And climate is not, is kind of something that definitely has been, you know, unfairly allocated in a lot of, in a lot of ways. And, you know, there's a strong case that climate is kind of being unfairly reallocated with, you know, global warming and the emissions that are that are happening and how that's kind of maybe good for Canada and Russia and Norway, but, you know, not very good for a huge number of other people. But climate is like the third one, right? So one is social network. Two is jurisdictional properties, and three is climate and

Balaji Srinivasan 1:09:24

the lives the laws and land

Vitalik Buterin 1:09:26

that Right, exactly.

Balaji Srinivasan 1:09:30

It's funny because, you know, one thing that when I talk about this with people, I think most people go immediately to how do you get the land or, okay, we need to write the constitution, right up laws. And I think that the people is actually the most important part because, you know, laws are boundary conditions, you know, in the sense of you're there. The things that are mandatory or forbidden, they're there the edge cases typically, you don't usually run into a law in your daily basis. You or like, as you're, as you're running around, and, and the land also, you know, it's hard to get land, and you might just start with the community, and then you can afford land rather than vice versa. So I'd always been, you know, that's a big premise of the narrative. I think those are two, I don't quite call them false doors. But their traditional ways that people have started things or thought about things that are perhaps a little less applicable today, the counter argument would be something like code is very similar to law. But you're writing code all the time. And I think the difference is, the machine doesn't do anything unless the code is telling you to do something. But the human does something unless the law tells him not to do something. And that's like, I think a difference in terms of code versus law. That's the difference between code. I know, I like less succinct code as law. But that is one difference in code law. Anyway, go ahead. You're saying so all those things climate you thought was or land was a big thing. So for those folks who have bad land, you would do x? What is x?

Vitalik Buterin 1:10:55

I think one way to look at it. One way to kind of separate out to what extent to what properties you need of jurisdiction is to think about the difference between starting a new country and starting a new city, right. So a city is also a social network. And cities are also very climate dependent. Right? So I don't know if you've seen the, the the charts and maps, but like, there's this really annoying fact about the US. And this is one of the reasons why so many people end up moving to California, is that California is like literally the only place in the mainland us that has what a lot of people would call a pleasant climate, right? So if you look at like number of hours of sunshine, California scores remarkably well. If you look at there is a statistic, number of pleasant days, like someone actually tried going through this, they defined the pleasant day as like a day where the temperature doesn't go out is somewhere outside of like, I think it's eight to 24 degrees. And the average might be like between 12 and 20, or something like that, and there's no rain, or there's no rain, and there's no snow on the ground, like California just like beats out of all the other parts of the mainland us there. Like there is a huge tax to being in California, right? Like there's, I mean, first of all, you know, regulatory issues and kind of, you know, the whole thing, this thing where like, pretty much everything that's based there needs to have, you know, some kind of a warning label saying that it's, you know, may contain traces of something that make it that may contain traces of causing cancer or whatever. But then there, the other one that I think is so much bigger for a lot of people's lives is the costs, right? And cost is California is expensive, real estate in California is expensive. And there even is this kind of pathological effects where, you know, the, the combination of building restrictions and property rights of rights, rules, and like a whole, a whole bunch of aspects of law basically means that like there is this you know, funnel where huge amounts of the just basically, this, the Silicon Valley startup ecosystem essentially gets really heavily, you know, a de facto kind of tax that's and it's kind of bloodstock to buy California Land voters

Balaji Srinivasan 1:13:29

with a lot of his deadweight loss. Also, though, because landlords would prefer to actually just build more. So it's like laws that are just causing them to jack up prices, because they can't build more units who prefer to do on volume, many of them anyway, go ahead. Yeah, you're saying

Vitalik Buterin 1:13:41

exactly right. But despite this, you know, the heavy drag on the yet doing things in California, there's like, it's still a huge network effects, right. And we've just been seeing how much of an uphill battle it is to, for projects to you know, real, relocate to other places. I've actually, I think, crypto, interestingly enough, has been possibly the one frontier tech industry over the last decade that has managed to really be really not be California centric, right. I mean, maybe there was a couple of others, like maybe, you know, bio as sub California and some Boston, and but I feel like crypto was like scores, crypto was scored better on decentralization than a lot of the others. But like aI very, very focused. You know, VR, from what I can tell seems pretty barrier focused. You know, Matt is there Apple is in Cupertino. You know, Google's that Google is there. Like it is a huge attractor. And part of that is just network effects and a strong initial network effect leading to even more network effects. But part of that is just like the climate being positive, right. So

Balaji Srinivasan 1:14:53

we're coming back up so so then how would you relate that to say to if you're starting a new country If you are addressing climate how,

Vitalik Buterin 1:15:02

right if you starting a new country or a new city, right? So this is where we have to kind of unpack things, right? Like one is that like climate is just an important property, right. So of, you know, whatever kind of in person physical things that you wants to start. And basically, if you're starting a new X, then the thing that you have to create from scratch is the network effects, right? Like, you're basically paying a large network effects penalty from the beginning. And there are people who I think are willing to pay the network effects penalty, because they want to different kinds of network effects. They're willing to sacrifice quality quantity of network to get quality of network, right. But the network effect is one thing that's being like, sacrificed or at least kind of radically reimagined out the door. As far as like what the climate needs are. That's, that's one of the two factors is going to judge where he wants to base yourself. And, you know, do you want to be in an existing country? Do you want to, you know, be in a big country? Do you want to be in a small country? Do you want to be in a particular part of the world? I think other things that I would add to climate, because like, even though they're not technically climate, but they're still kind of features that are given to us by the physical world, time zones are a big one. Right? So if you've watched, like a remote work for American companies, it's much easier to be in Argentina than it is to be in a Europe despite those two being equal distances from America, because, you know, in Argentina, you have and in Latin America, in general, you have timezone compatibility, right. So there's a lot of those deciding factors that come from the natural environments. And then there are also these deciding factors that come from the lock, right? And so, basically, the question is, like, what, what is the benefit that you're trying to achieve? And what are the things that you're willing to sacrifice? A lot of people don't really have that much fancy stuff that they wanted, that they need to do in terms of in terms of law, for example, right, like for a lot of manual people in projects, actually, the big law the that people don't sometimes don't realize that is a law, but that the law was through that they wants to escape is a US immigration law, right. Like the US and a lot of other wealthy countries, they do have middle quite restrictive, immigration, a lot of ways and like, The tragedy is that sometimes it's pathologically restrictive and that, like people who, you know, try to go in the front door can you know, apply for visa and all that, like in India? I think it takes like, what, 900 days to get an appointment or something crazy like that. But on the other hand, you know, if you're willing to take the the route of, you know, jumping fences a bunch of times, then yeah, like, sometimes that's even easier than, you know, taking

Balaji Srinivasan 1:17:58

a narco tyrannical in the sense of like the law isn't enforced on the law abiding, or the law is enforced, but only on law abiding, and is not a forced on, you know, it seems so illogical. But it's not it's like those who are habitually the law abiding, or I have the book thrown at them for small immigration things.

Vitalik Buterin 1:18:16

Yeah, but but like US immigration, why like is like, it really is a huge drag on, I think, a lot of difference. It's

Balaji Srinivasan 1:18:24

also not letting me on the high end talent. It's basically just deadweight loss across the whole system. It's not it's not smart.

Vitalik Buterin 1:18:29

Yeah. Right. And it's one of like, it's one of those things that's like, potentially Asia, like a 2x, or 3x tax on the many groups of people's ability to operate, even if they're doing some completely normal thing.

Balaji Srinivasan 1:18:43

And the thing is, it's one of those things where basically, different groups, it's not simply about making immigration law harsher, as some people think it is. Because what happens is that it's just those laws are just enforced against those people who are, who are not the folks that are. It sounds like the guy who came across and shot somebody, you know, it's a person who's filling out paperwork to be a programmer. But anyway, so we kind of got off track a little bit here. So climate of course, I mean, the one thing about climate just for a second or land is, I would argue, potentially, at least for a new country, the more desirable the land, the less desirable the land. And the reason I say that is the more desirable the land, the more contentious it is for people to try and get it whereas if it's something that's godforsaken middle of nowhere, like wood burning man was built on, it's easier to build on it if your goal is freedom, as opposed to a pleasant climate. pleasant climate, I think is like the very hardest thing to get because it's so broadly appreciated. And so mimetic in a sense, so you might have to suffer a bit, you know, land wise in order to, you know, get freedom.

Vitalik Buterin 1:19:51

Yeah, I think that's, I think that's very possible, I think, but I think the bigger lesson from all of this right is that there are ways to With improve your situation in terms of law, right? So like, for example, there are lots of countries that are visa free to Chinese people, there's an unfortunately smaller number, but definitely still a significant number of countries that are even visa free to India people. I mean, India itself visa free to India people, obviously. But at the end, then there are countries that are just easier to get visas for, right, like there's, you know, countries that have visa regimes, but, you know, they're not, Indians are not going to have to wait 900 days to get one. But at the same time, like if you have to be willing to sacrifice something, right, like, the hard thing in the world, is to find kind of like actual Alpha like to fun to actually find something which is like, further along the frontier along all the dimensions that people care about

Balaji Srinivasan 1:20:50

just pre to optimal of some kind you're trading off a for be, exactly, the easier

Vitalik Buterin 1:20:55

thing to do is to try to is to get a better score on dimensions that you care more about, by sacrificing something on dimensions that you care less about, either, because you have like, you're you just are the kind of person or you're more or you have some trick for like tolerating a lower quality on those dimensions, or because you have like some different path for getting those same things, right. And I think the big thing, like one of the ways to think about the like network, the start of a new city or starting a new country genre, is that it's basically the genre of saying, Well, what are the things that you can get if you're willing to completely sacrifice on network effects? Because you're gonna bring your own network effect? Right? If you can bring your own like, there was a lot of, you get a lot of options, right. So, for example, over the last year and a quarter, one of the things that I do this, I intentionally spent more time than I usually do in like very small towns, right?

Balaji Srinivasan 1:21:53

Tell Tell everybody about it, because most people don't know about. Okay, yeah, so,

Vitalik Buterin 1:21:57

Svalbard is an island, 1000 kilometers north of Norway. It is officially part of Norway. But there are international treaties that have existed for about 100 years that say that anyone from the world is legally allowed to come and love their

Balaji Srinivasan 1:22:12

open borders act, but it's so forbidding that many people don't actually exercise that right.

Vitalik Buterin 1:22:17

Exactly. So VHL largest city in Svalbard is a town called one year Vienna does a population of about 3000, it's a longitude is 78 degrees north. But interestingly enough, the climate is somewhat less inhospitable than it seems right. Because basically, that entire region, like all of Scandinavia, benefits hugely from the Gulf Stream, and the Gulf Stream, it makes all of those regions significantly warmer than like lots of stuff at comparable latitude. If you'll look at, say Trump, so a city in Norway, 69 degrees latitude, the lowest temperatures, there are like actually not much worse than the lowest temperatures in Toronto, which has 45 degrees latitude, but the latitude there is much higher and then Svalbard, like what I was there in September, the temperature was like between minus five and plus five. So like, actually, you know, Yorkers

Balaji Srinivasan 1:23:08

can do it minus five and plus five. I guess that's a doubt at least it's not like, too bad,

Vitalik Buterin 1:23:14

right? I mean, it does, it definitely does get bad. And, like, in January, for example, right, but so, you know, there is a reason it hasn't grown to more than 3000 people, but it's like, it's less bad than a lot of people think, right. And there is a town of 3000 people. And that's how it has a large supermarket, and has a post office, it has a hospital, it has an airport, and it has one of the best sushi restaurants I've been to in a while.

Balaji Srinivasan 1:23:42

Okay, now it's gonna have lots and lots of people visiting, they're gonna have 100,000 people.

Vitalik Buterin 1:23:49

Right? Like, I think marginally, it'll medicine, we're getting more people coming there, right. But it's definitely, you know, encourage people to just like go and visit and you know, if they can visit for a longer time, like, you know, give them up to two weeks or so just to get a feel for like, you know, what it feels like to actually live in this kind of place. But it's interesting how even in a town of 3000 people like that kind of infrastructure you can get right, like you have a supermarket, you have a hospital, you have sushi. If you do the math, then 3000 People basically means that you have about 30 people per a year level, which means that so long European also has a school, right? And at three people, that's enough for like one or two classes per grid. So it's actually surprising like how few people need enough network effects to at least have the goal is always to have the boring basics, right? But the one thing that you can't get with 3000 people is you can't get like people that are organized around particular interests, right? Whether that's for like work or whether that's for like for for you know, just like social interaction reasons, right.

Balaji Srinivasan 1:24:57

It's often talked about how the early American call Use only about two and a half million people at the time of independence, something like that, you know, the Greeks very small population at the time, I think Rome was only like a million people. Yeah, or something. Like, I mean, these are much smaller than we think of, in Rome, the entire empire. But I think the city of Rome, in the Roman Empire was like a million people. At that time. I might be wrong, but I will check that in the show notes.

Vitalik Buterin 1:25:22

Yeah, but the thing that's fascinating, right, is that 3000 people is like, actually, yeah, it's totally viable from that standpoint. Like I think, as a residential, we 1000 persons out and your main risk is getting bored, right? And it also you have a risk of, you know, not having not being able to find a good job. And one way to answer that is to basically say, we're only going to cater to remote workers and remote work is only going to grow from here. Another way to answer that is basically, through the same channel that we can answer the being bored problem, which is, well, let's look at the one type of town that already exists that has a small population, that is still very interesting for people to often spend large amounts of time and that's the university town. Right. So now we have something like Ithaca, in upstate New York, population 30,000, to home of Cornell. And, you know, if you care about any of the academic topics that Cornell specializes in, then like, you know, you're, you're not going to be bored, right? And so if you, like, really concentrated, you really tried to create a network effects around specific kinds of people, then you don't you don't even need 30,000, like even 3000 is going to be enough, right? Like if half of those 3000s are people who are into, let's say, crypto, or let's say a longevity research, you can create an environment that's like, gives people enough of that, that it with a surprisingly small, a smaller size, right? So then, you know, what do you get for that? Right? One of the things that you get is, I think cost savings, right? Another thing that you get is like the right type of network effect, right? So you are a different type of network effect, you sacrifice on quality, and, or you sacrifice on quantity, and you get quality. Cost savings are also huge, right? Like, that's one of the big things that I think has happened over the last 10 years is that prices have increased in the US significantly more than they have increased. And a lot of other

Balaji Srinivasan 1:27:30

you're talking about this graph of that. I'd love to I think that's true, but I'd love to see that. I agree. But go ahead.

Vitalik Buterin 1:27:36

You know, I that's actually something that I haven't checked yet. Right. But like it just like it feels true. From my from my personal experience in even Switzerland, I think the price of like tea and Starbucks has increased very letter little or like maybe even not at all over the over the last 10 years, which I thought was fascinating. Right? So costs are huge. And the cost difference between being in the US and being in basically any other country is like getting to be pretty significant at this point. Well, I think to be fair, it's always been significance. But I think what we're seeing now is, it's qualitatively different the same time the end you can earn outside, right, but it's like at the same time, right? A significant right. One thing is that a significant cost difference is staying. And other thing is that now you can earn outside and more generally, like the quality of life difference between being in a major center and ads being in you know, what gets called a kind of, quote, periphery location,

Balaji Srinivasan 1:28:34

the cost of increases the benefits of decrees. And so they're coming closer together. Yeah, exactly. Right. So

Vitalik Buterin 1:28:39

that creates a big opportunity. Right? So cost is one. And then you know, we've mentioned law is another one. And that's an interesting one, because I think it's, it's one that a lot of people don't think about it. And it's easier to not think about, if you're in a situation that fit that is fine, like that feels to you as being normal, right? Because if you're just an average American, and you know, you're you're doing some completely Norby industry, that's, you're not probably going to feel, you know, law as like, as being a large drag. And it's, and it's even easy for you to get into the mindset of believing that, oh, it would if someone does come up to them and tell them the law is a big drag. Well, possibly, they're doing something wrong, and they have something to hide, right. And what's interesting is like, the further away you get from like that, kind of, like privileged, rich country context, the and like, I think it's even important to kind of talk about the rich country part, right? Because, like even the privileged discourse that exists with other rich countries, it tends to, like, at least in my view, it focused a lot on underprivileged people within those rich countries. But there's also this entire map like much larger Lots of people are in the world, around the world that are not even in those rich countries, right. And I think even the US and even progressives in the US often still have, you know, what's de facto very much and out of sight, out of mind mentality towards a lot of those people. But the further away you get from, you know, being in, in a Asia citizen of a rich country, the further away you get from being in a kind of average industry, the more you start feeling these kinds of things, right. So, obviously, the crypto space, I think, has felt regulatory pressure is very keenly from the, from the beginning. But in 2022, I think caring about the law, like really universalized, in a lot of ways, right, in the sense that, like, there were a lot of countries, like a lot of non US countries and even, you know, anti Western countries that have historically been authoritarian, but the political geopolitical environment for the last 20 years has been this kind of fairly stable one, and fairly friendly one where even in the authoritarian once the authoritarianism kind of felt theoretical, and it's like, oh, you know, if you personally aren't somebody who like really cares about criticizing the government, and like, why would you even do that when I'll just focus on your family, then you're not really feeling much right, you know, a Yeah, you know, focus on the family, you know, enjoy the enjoy the local culture, enjoy the trains, and it doesn't look that bad in the normal case, from like, that kind of cultural standpoints. But like in 2022, in a whole bunch of ways, the authoritarianism became real, right. And it became something that even really heavily touched the lives of normal people. And it's something that just overnight created a whole bunch of like, basically homeless nomads, right. Like I personally have multiple friends who, in 2022, it just like, packed up left their country, and they're just jumping around. So they're not really clear where they live. And these, in a lot of those cases, these are even people who were not that worried about the political environments of where they came from before 2022. It's right. So on the one hand, you have

Balaji Srinivasan 1:32:18

something called visual analogy that you may or may not agree with, this is like a, you know, for example, when you're driving on the highway, your speed limit could be minimum 40 miles, maximum 55 miles per hour. And that's like a, you know, it's a bounding box that you have to be within your v has to be within that, right. And you start extending that to every other set of laws. And essentially, regulations in the law are a bounding box, where you have to remain within that right speed greater than this, you know, not less than this, your, you know, bank account balance, at a certain time can't be over this or under this are in that seven out of, for example, building cannot be more than this high, you probably cannot be more than this low, that kind of stuff, right? Multiply that across 1000s, hundreds of 1000s of laws, all the laws, so you have to sit within this box, and most the time, most your environment, what's your life is configured to live within that box. So you're so in the box, that you don't even feel the edges. But then sometimes, like in 2022, or really recently, those edges suddenly like move, and they kind of put you outside of that box where what was totally normal suddenly becomes abnormal. And you're, you know, looking around, you're like, how did that happen? I thought it was just totally in the mainstream. And then this just sliding of walls got you outside the mainstream. And I think that in many ways, what we're we do in crypto is for the powerless and the power user in different ways. They're pushing the limits, like the power list wants to hold on to a bank account. And the power user is pushing the limits of what a bank account can even do. And sometimes they're the same person or the same entity, like, like by Nance is both absolutely a power user of money. But I'm sure early on, maybe now, but you know, they had, like it's real, it's called regulatory issues, or what have you, right, like they, you know, they will report it. You know, I have no beef with finance whatsoever. I'm just saying, sometimes it's the same person the same time, those are not necessarily disjoint categories. And so that reminds me of what you're saying, which is, these walls on this box that people took for being static, became dynamic, these constants became variables.

Vitalik Buterin 1:34:27

Right, like so like, I think, up until 2000, that and that's when the two of you just like had to be either one of those people who situation was that you were in a beside a wall, we're in a court or in a corner of the room, or you just had to be an explorer that will definitely touch against them. But like, Yes, since 2022, the walls moved. Yes,

Balaji Srinivasan 1:34:46

exactly. And so, so I'm bringing that back to the start a new country thing, right. So if you could start a new country, how would you say, you know, we talked about the climate aspect. So when we talked about cost, so and now we're talking about law, so it seems cuz like when you try to do is hyper deflate cost? Or, you know, obviously I know I'm poking on this, because I've thought about this a lot and you know, but you know, what is it? Would it be Aetherium? Based? I assume either Landia, right? Is there? Is there some funny Etheria? I don't know, what would be the would be the land, if the dollar is the currency? Well, it doesn't actually conjugate quite that way. Right? You might have Germany and the Germans, but the Mark doesn't quite conjugate. So let's say you've got a theory and what is the theory on land look like? Right. And is that related to this?

Vitalik Buterin 1:35:35

Yeah. Well, right, it's a good question. I mean, I think one of the questions is like, what actually makes sense to create as a UA and then the other question is, what level of tie it should have to do with the area? Right? Because I think on the second question, like, one of the things that I do wants to be careful about is like, not assuming, you know, more agreement between people that actually exists. And, you know, not basically, like, I want to avoid getting into the, like, intersectional activism trap, where, you know, you start by creating a community around one cause, and then you start kind of like, adding on more and more of these coalition partners that have their own causes that, you know, they're, they're good at expressing in your language, but then a lot of the time when you poke into the DCLs, you're just ends up getting stuck in a mindset where, like, the, the, you know, the intersection of all the causes, actually only

Balaji Srinivasan 1:36:30

appeals against 1% people are who are against what you're against, but they're not for what you're for. That's the tricky part where you will win the war, but you won't win the peace.

Vitalik Buterin 1:36:38

Go ahead, right? That's exactly right. So, yeah, to what extent like some, you know, new city or your country started by theory, and people like should even market itself as being a part of the Ethereum verse or even being a central part of the Ethereum verse, right? Like there were stewards differential like, like, to what extent? Like I think it would be, it might even be good to be careful about creating the impression that if you're on a theory in person, you have to like this other thing. Like, I think it's

Balaji Srinivasan 1:37:10

big enough. It's millions of people using it. So

Vitalik Buterin 1:37:14

Exactly. Like I think there's even, you know, virtue in kind of, you know, people who are in multiple movements, like basically, yeah, say, like, saying that it's okay to be with me on one thing, but not, but not the other thing. On the second question, right. I think the Yeah, the answer is maybe no. But at the same time, I do think that there's a huge momentum, the momentum within the Etherium ecosystem for wanting this kind of thing. And so I think, like, the crypto world, and the AGI, and the theorem world are definitely like natural places to start and I think are just that inevitably going to be big parts of, you know, any new country, that's or, or any new city that gets created, right? So

Balaji Srinivasan 1:37:59

it's, you know, there's about 300 million, depending how you count it, but like Coinbase, and binance, alone are a good chunk of at least 300 million crypto users worldwide who hold at least some cryptocurrency. And I kind of do think of them as an Intuit country, because, you know, they, they all need to kind of accept each other's coin, right? And so they have to have in a sense of bilateral relationship, where it started kind of currency and network first. So it is a good group of people, early adopters, somewhat risk tolerant that could pioneer something like this, even if they aren't all, pure 100% Aetherium, or what have you, they likely certainly know about it and wholesome. So I think it's not a bad analogy, or at least it's a it's a predicate for it. It's not it's not the full thing.

Vitalik Buterin 1:38:45

Right. I mean, I think one good analogy for this might be like, so if you'd like forecaster, for example, right? So this actually gets to another topic that I think is important, which is that one thing that I think network states and social networks have in common, is that especially looked at just to be able to get them started. It's not about the institutions. It's not about the technology. It's about the people, right? Like if if you imagine five clones of forecaster, where let's say one of them is run by Dan Romero, and other of them is run by let's say, a American trans social justice activist. A third is run by let's say an Indian person. A fourth is run by let's say, yeah, some Eastern European Ultra nationalist and a who by the way, have likes Bitcoin and thinks that Bitcoin is the only coin and a fifth is run by like some 14 year old in in huncho. That's just that happens to be a get interested in in the Let's say you like West African philosophy and has some, you know really crazy yet a fusion ideology that for whatever reason they can convince such he can convince a significant fraction of Chinese people to or Chinese Generation Z or generation at once offs but no one else, right like those five forecasters are going to look very different

Balaji Srinivasan 1:40:20

I've ever concrete a concrete example for you don't share chat is it is that third example it's an Indian social network that is non English. And so there's a Hindi channel of Bengali, generally Malayalam channel, Canada channel and so on, right Donald jail. And so that's an example of something which is Twitter like, but looks very different than actually because forecasts are as Ethereum functionality built into it. So therefore, it's like their Twitter. This has it's just partitioning different axes, you wouldn't have thought of, okay, which Indian language do you speak as being a primary axis for forecaster, but it is for this, right? Whereas forecasters, like which NF T's you hold? So it's just like moving on a different axis? Go ahead.

Vitalik Buterin 1:41:04

Yeah, right. But I think the larger point, right, is that like the, the root community is important. And I think this was a lesson that I think I kind of intuited for Aetherium itself, right? That's, like the different one of the differences between like web two, and web three, I think, is that like things like Facebook, they just tried to scale the world and reach out to everyone, right? And things in the crypto space and things in the network state space and things in? Even, let's say the radical democracy is based like they're, they are more rooted in specific communities. And they do they are, I think, kind of more respectful of the idea that like, well, sometimes network effects aren't positive network effects are negative, right? There are people who are negative value to other participants, especially when you create like, products or ecosystems where the interaction between the participant says yes,

Balaji Srinivasan 1:41:58

I think this is like saying, like law is meant for virtuous people. And, you know, like, I'm paraphrasing, but that's like the early, the Founding Fathers said, you know, we can write all this stuff down, but it's meant to deal with things in the breach, you have to basically be good, you know, or like, you have to be virtuous. So it's kind of like the structure is something that expresses the community's will. Okay, so basically, we've got about, you know, 15 ish, 10 minutes left. And so just to recap, what did we talk about? Well, we talked about a brief history of DRM from concept to the present day, we covered the merge and the Dow hack, and OFAC and privacy and a bunch of other things. We talked about, you know, what new country you'd start and addressing problems like climate and cost, and whether Ethereum people would be there, he talked about actually the really interesting place Svalbard anybody can go to, and how what open borders would actually look like in real life we have, we have an empirical example of it should be on the cover of a bunch of books, maybe it'll, it'll be changed by that attention. And I wanted to maybe touch on one last thing, which is, you know, we've talked about how to improve the world, we've talked about how to improve Etherium. How can people improve themselves? Which they read, watch, and listen, what online courses? What should they be doing? You know, like, how, what would you advise people to do to level up?

Vitalik Buterin 1:43:23

Sure. So I think one of the reasons why this is an interesting topic is that it's not even entirely separate from the network's state topic, right? Because, to a very surprising extent, I think, like Learning and Motivation are very social, right? And they like, being really successful in those things actually, depends on having the right social environments. And I think this is one of the key reasons why, like, for example, you know, of course, era, the various MOOCs have failed to completely replace Stanford, to the, if anything, like the extent that a lot of people were predicting, like, I mean, one aspect is definitely like, just like Credentialism, and the fact that the books don't give you a fancy diploma, but I think the, like, the social elements to Learning and Motivation is the other really big one. And actually, it's like the question of like, what new country? Would I start, like, that might even be part of the answer, right? Like, I, I personally, would start with charity before our country, by the way, I mean, I would like a city located a city located in a country that has enough of the right, you know, law properties to get the ad that you need 20 of that, right. But like, like trying to, like actually, you know, tweak the route of, you know, over the legal system and like, go at the sovereign level, like that's both. It's a very difficult and it's a very challenging project. And it's just like all kinds of international diplomatic risks that you don't really want to get into unless you're fairly big, right? And so, I think You know, working with existing countries is just something that's going to realistically make sense for projects going, I think going into the future for a while, right. But starting in your city in but and at the same time, like having the country aspects to the extent that you're creating a network effect in a country, that that would otherwise not have that network effect. But one of the things that I would add in there is definitely, like, have a stronger emphasis on like, education as a part of the culture, right? Like, imagine going through your entire life, like your living experience for your entire life, being a university at 10 at 10% intensity, right, so not a university 100% intensity, but you know, keep it on 10% 10% intensity, have like, very easy and like very socially available opportunities to constantly keep learning new things,

Balaji Srinivasan 1:45:55

and how you say this, because it basically sorry, I mean, to cut you off, but like the, it's like buying a car, a college is just like buying a car and not budgeting for maintenance over your life, the entire cost is up front. And then you have no gas or oil change. And then people are like, Oh, I studied him, like, study for four years and doesn't remember any of it is extremely expensive install of ostensibly education that is, has zero maintenance associated with it, you know, don't me Go ahead. So I'm agreeing with you.

Vitalik Buterin 1:46:20

Learning that I think, especially in the fast changing world that we're in like it really is something that should be a lifelong process, right. It's like and especially once we you know, solve longevity, I think I found that either Aubrey de Grey or someone else talked about this that like historically, the pattern has been steady work retire, but the new pattern might be steady work, retire, steady work, retired, steady work, steady, retire, work, retire steady, work steady. But the other approach, of course, is still like really integrate that you have more than that, like studying should not even necessarily Yeah, BFA as it should just be a regular part of your life. Right. But so the social aspect is important, right? And I think, like one of my number one advices, to like, probably young people and all people everywhere is to be intentional about the social environments, and, you know, to have to have friends that help you bring out the virtues that you want to see in yourself is something that, like, I just think that's so hugely important, right? Like, you're, there's just all kinds of subconscious ways in which your behavior changes for the better when you're around people.

Balaji Srinivasan 1:47:29

It's funny, because the way to prove that to people is just actually AI prompts, like the prompt is so insanely important. And so you're just getting effectively prompted in a sensory fashion by Oh, okay. Yeah. All right. They're lifting weights, I should probably go and do that too. Like, as a small example, or oh, they're setting Matt Oh, yeah, let me take a look at that. So like, the prompts that you surround yourself with both digitally and physically have like, such an influence on you and that, and one of them is like, your social network. It's such an obvious thing, but people don't usually think about it.

Vitalik Buterin 1:47:59

Exactly. Yeah. But I think, you know, until you can, you know, until we have multiple of these established, intentional communities, and all of this just whether they're, you know, creating new cities, or UK or new countries, or, you know, you Bitcoin monasteries, or, you know, whatever people in that camp or trying to create or whatever, it's things that you can do while you're by yourself. You know, I think, like, motivation as the number one problem is, I think, one of the key lessons that I've learned, and there are, like, less, not social ways to get that too, right. So like, one of my personal experiences, right, is that like, I've used Duolingo for a while, I mean, I've used that as one of my ways to kind of practice on my Yes, Spanish show for a while, have also experimented with some of the other East Asian languages with it, too. And it's, like, the way that the app prods you into like, just trying to like, into doing at least one thing every day, and the way that the app makes you, like, even feel the fear of, you know, losing a multi 100 Day streak, if you decide to go away from it. Like, it's stupid, and it's like, so obviously, you know, gamified in a way that even seems a little bit trashy, but like, for me, it works somehow, like as though there are limits to it, right. So like, one of the things that logistic kind of goes down the path of Duolingo in particular, that I found that other people have found is that it pushes you to optimize, not for the game of knowing the language, but for the game of winning Duolingo, right. Like for example, sometimes Duolingo gives you a task like translate this sentence, and like DeVivo 20% of the time, you don't even need to look at what's at the top of the screen like the yes sentence in Spanish or whatever you just like go down you look at the list of words that they give you to select from. And there just obviously is one sentence that you need. English in English you can construct from those words. And so you look at the bottom of the screen, you just like tap that sentence and you're done. Right? And like you've won the game, but you haven't for actually learned any Spanish. Yeah, so I do think it's like importance to mix other tools, I mean, I could actually tell you the path that I took to learn Chinese. So. So at the beginning, I just started with Pimsleur, that's p i, m s, l, EUR. It's the set of online course, or basically audio courses, right. So it's like a 30 minutes, like, based course that they have 90 of them read, you go through them, and it gets you up to a very basic level. Right? So I think different people might have different experiences. But for me, personally, you know, there's this kind of, on, there's a stereotype that exist online that like, it's possible to somehow pick up languages just by like, you know, going into a country, and eventually, you kind of pick up words, and obviously, two year old kids can do it, and possibly extroverts can do it, but I totally can't say, right, like, if you just like stuck in the into China, with no training, possibly, after even after 20 years, I would not like, you know, quote, pick up Chinese from zero would just be miserable. And, but, so for me, personally, you know, for the first stage of learning languages, like, you know, you need the structure and stuff. So I use pandas, or I also used some of flashcard apps to warn the characters. And then once you get above a certain point, there's like, a certain level of like, medium comprehension, we are, that's the point where you can start picking the language out from right, so before medium, you know, you start with the, you know, the podcasts, I did pimps, or I think I did I Chinese pirate and a couple of others. Once you get past that point with these flashcard apps, then you go look like for me, personally, I you know, and I actually spent like, a total of about one and a half years in China, and you know, sometime in Taiwan, and you know, sometimes it's in Singapore, and just like a lot of time talking to just like other Chinese friends that I heard outside of China. And it's once you get to that level, then you actually can just like keep learning by by just talking to people, right? So talking to people. And this is, ironically enough, actually, there is one particular way in which Chinese is easier than languages. Like, you know, German, for example, which is that if you try to speak in German in English, or history, try to speak in broken German. So someone in Germany, very often have a theory quickly switched to English.

Balaji Srinivasan 1:52:40

So you don't get those training wheels in China. Like this might pop.

Vitalik Buterin 1:52:45

Exactly. You don't get the training wheels, right. But like in Chinese Oh, no, no, no, you know, there's lots of people who barely speak English at all, and like, they're, they'll be happy to try their best to understand you. I'm actually Spanish as well. It's like, I think in maybe in Spain, a bit more English speakers but in like deep Latin America, you know, definitely lots of people whose English is very bad. And you can definitely get your Spanish up there, especially if, if you try recollect past. The other thing that I did in China is like you just to like walk around the city. And you will look at street signs and the street signs, they all have the name of the street and Chinese characters, and then the name of the street and pinion, right, that gives you the pronunciation. And so you just like, look at street names, and you passively learn a whole bunch of characters just like as part of your life. Right? So like, basically, yeah, so to get from like zero to med, you know, you do the podcasts that you would do the flashcard apps, you do the like very structured courses, and then to get from mid to high, that's when like picking things up and, and the more social approach.

Balaji Srinivasan 1:53:49

That's a good combination. So basically, if you're an adult to summarize, if you're an adult, don't just start with immersion, do courses to bootstrap yourself to that basic level, get your BIOS up and running. And then you can kind of assimilate content. If you're two years old, you might just be able to be dumped in, right.

Vitalik Buterin 1:54:05

Exactly. Yeah, I said for technology, that for like things like cryptography, for example. I one of the main lessons has been the importance of learning by doing right like if you just, you know, if you just read and listen, there's just it's just so easy to kind of go you know, in one year or, or one eye and out the other. But if you actually make yourself build the thing, that's the way that like, you actually kind of get the knowledge into muscle memory like you you work with the concepts with your brain in a much deeper way. You're gonna

Balaji Srinivasan 1:54:39

schaums Outlands it's basically like I've very similar shapes, things are pretty famous in India, basically are these yellow outline books, where they just have solid problems. They just go straight to solve problems. And so you just go straight to the exercises. And it's like relatively minimal introduction. Just go straight to exercise, just like just start working out, not reading about working out you work out, you know And that's similar to this, where basically by coding, coding it by building it and then trying to break it, you understand it in a way that just theory alone reading about it doesn't get you something

Vitalik Buterin 1:55:09

that's actually fascinating. Like, if it's true, I tried to the idea of teaching people even without contents by but by by just the like repeatedly giving them exercises that are like one inference step away from the previous exercise that they solved. Regardless of you know, what specific tool you use, I think, if you weren't, if you want to learn about snarks and build a snark, if you wants to learn about how Aetherium works, build an Aetherium aplique implementation and get it to pass some of the tests or may even get to a point where it passes all the tests, then, if you want to learn about like homomorphic encryption, build an implementation of amorphic encryption, right, but if you want to learn about AI, then like actually, you know, build a neural network that's, you know, at least powerful enough to do some, like basic job of recognizing digits or whatever. Right. That's, I think that's one of the things that helped me. In 2011, I went through it some the machine learning course on I forget where the the, there was one in Udacity, and one on Coursera. That was one by one was by Andrew, right? Well, that's my special it's run one was by Andrew Yang. And they were very hands on right where, like, you had to actually implements you know, a lot of like, things like a Star Search, search algorithms and like you have all that, you know, multi layer neural networks and all of those things. So actually implementing a thing and just like gives you this a very deep and hands a hands on understanding about like, what the thing is, and you know, how to work with it, it just gets intuitions that you can't get by just reading a textbook. So highly encourage taking an experiential approach. You know, definitely, it's the No, I think there it's the equivalent of like the Star Wars, like kind of Jedi tradition where you're supposed to build your own lightsaber. Right? It's do or do not there is no try except for the Merkel patrician tribe, which by the way you should implement, that's a good one. That's a good so lesson, you know, build your own of people, laws and, and the land, you know, you sacrifice quantity of network effects, go for quality, you know, get the, get the right group of people, you know, be modest on online, especially, especially at first, you know, provide focus on the public goods motivation. If you want to learn about things, then, you know, go and build your own lightsaber, and I think awesome, lots of stuff for people to do.

Balaji Srinivasan 1:57:45

Thanks for coming on Vitalik, and this is the first episode of The Network State Podcast.