Bankless - 46 - Vitalik Buterin on Why Proof of Stake?

Episode Date: January 4, 2021

🚀 SUBSCRIBE TO NEWSLETTER: http://bankless.substack.com/ ✊ STARTING GUIDE BANKLESS: https://bit.ly/37Q17uI❤️ JOIN PRIVATE DISCORD: https://bit.ly/2UVI10O🎙️ SUBSCRIBE TO PODCAST: http://podcast.banklesshq.com/ 👕 BUY BANKLESS TEE: https://merch.banklesshq.com/ -----GO BANKLESS WITH THESE SPONSOR TOOLS:  ⭐️ AAVE - BORROW OR LEND YOUR ASSETS https://bankless.cc/aave 🚀 GEMINI - MOST TRUSTED EXCHANGE AND ONRAMP https://bankless.cc/go-gemini 💳 MONOLITH - GET THE HOLY GRAIL OF BANKLESS VISA CARDS https://bankless.cc/monolith 📈 KWENTA | DEVIRATIVES TRADING WITH INFINITE LIQUIDITY https://bankless.cc/kwenta ------Vitalik Buterin on Why Proof of Stake? Vitalik Buterin comes to the Bankless YouTube to discuss a number of topics surrounding Proof of Stake! We begin discussing the Beacon Chain launch, and the Ethereum's road so far up to this point.  We also discuss the theory and thought-process behind Proof of Stake! Vitalik recently wrote a blog piece 'Why Proof of Stake', and we walk through this piece with him! Why Proof of Stake: https://vitalik.ca/general/2020/11/06/pos2020.html Lastly, we discuss one of Vitalik's blog pieces 'Concave and Convex Dispositions, and how these different personality types characterize Bitcoin and Ethereum, as well as the design of social groups and human governance organizations.  Convex and Concave Dispositionshttps://vitalik.ca/general/2020/11/08/concave.html Other pieces mentioned: Trust Modelshttps://vitalik.ca/general/2020/08/20/trust.html ------Don't stop at the video! Subscribe to the Bankless newsletter programhttp://bankless.substack.com/ Visit the official Bankless website for resourceshttp://banklesshq.com/ Follow Bankless on Twitterhttps://twitter.com/BanklessHQ Follow Ryan on Twitterhttps://twitter.com/ryansadams Follow David on Twitterhttps://twitter.com/TrustlessState -----Not financial or tax advice. This channel is strictly educational and is not investment advice or a solicitation to buy or sell any assets or to make any financial decisions. This video is not tax advice. Talk to your accountant. Do your own research. Disclosure. From time-to-time we may add links in this channel to products we use. We may receive commission if you make a purchase through one of these links. We'll always disclose when this is the case

Welcome to bankless, where we explore the frontier internet money and internet finance. This is how to get started, how to get better, and how to front front run the opportunity. This is Ryan Sean Adams. I'm here with David Hoffman, and we're here to help you become more bankless. David, we have a fantastic episode with Vitalik. What do we walk through with Vitalik today? Yeah, a number of different topics. We wanted to get Vitalik on the podcast on the YouTube to go through some of the current events in Ethereum as he is digesting them, as we are digesting them, and also go through some of the blog posts that he has written on his website, Vitalik.ca. And so we begin talking about this conversation of the beacon chain launch, what that was like, what that felt like, just the

emotional response from Vitalik. But then we go into his article, Why Proof of State, which he wrote in November 2020, a really important article to understand the arguments behind why proof of stake is so valuable, why it's uniquely compelling and why it's security assumptions seem to be really, really strong. And that is something that we are going to see being tested in 2021. And then we finish off with a conversation of convex and concave dispositions, which Vitalik characterizes people as either being concave or convex. And we'll get into what that means in the podcast. But he, and I think listening to all of these different topics, it's a diverse podcast with a number of different topics.

But framing it in a convex versus concave disposition, I think adds an extra flavor or actual layer to really appreciate as to how people will come to argue about why proof of stake versus proof of work or why Bitcoin versus Ethereum, why expressivity at the base layer via or trying to keep the base layer extremely concise. So it's a wide range of conversation,

but there's still a through line all the way through. Yeah, I feel like that through line is like culture. right? It's like philosophy. These two communities, Bitcoin and Ethereum had different schools of thought, different philosophies. And that's why they have made the decisions that they've made. And that comes across, of course, with the launch of the beacon chain. And then also this conversation of why proof of stake is better than proof of work. So it's a super interesting conversation. We almost start with like the manifestation of the value system in the culture and then work backward to the philosophy. kind of the question of why. Vitalik is always a pleasure to have on. He is like a super deep and

insightful thinker. And he's not just technical, but he's almost like an anthropologist. Some of the ways he thinks about the human and social element of these crypto systems. And I think that's what makes all of this unique. One note, of course, we did release these episodes on YouTube in November and into December, but this is the first time we are publishing them all in one place on the podcast for your viewing pleasure. And of course, guys, you can always watch the full video interview on our YouTube. That's utah, YouTube.com slash C slash bankless where you will find that. All right, guys, we're going to get into the interview with Vitalik Buterin. But first, we're going to take a moment to talk about some of these fantastic sponsors that make this show possible.

Gemini is the world's most trusted cryptocurrency exchange. I've been a customer of Gemini since I first got back into crypto back in 2017, and it has been my main exchange of choice to make my crypto buys and sells. Gemini is available in all 50 states and over 50 countries worldwide, and on Gemini, there are markets for over 30 various crypto assets, including many of the hot defy tokens like Wi-Fi, Avey, Uni, and also they are one of the few exchanges that has liquid dye markets.

Having both the option of logging into the Gemini.com website or instead opening the Gemini mobile app has allowed me to be able to access any and all exchange and on or off-ramp services that I've needed to on a moment's notice. With instant deposits and fast withdrawals, I'm able to make my money do the things I want it to when I want it to. You can buy crypto safely and securely on Gemini

with the peace of mind of knowing that your investments are insured and protected with industry-leading cybersecurity. You can open up a free account in under three minutes at Gemini.com slash go bankless. And if you trade more than $100 within the first 30 days after sign up, you'll be gifted a free $15 bonus. Check them out, Gemini.com slash go bankless. Synthetics is Ethereum's decentralized derivatives liquidity protocol. What does that mean? Synthetics is a platform for creating and trading synthetic assets, which are assets that are priced via an Oracle rather than bids or asks. Traders can use the quote.

Quenta Exchange, which hosts and trades all of the synthetic assets created by synthetics. Traders on Quenta can trade synthetic tokens like SBTC, S-Oil, or S-D-Fi. Because Quenta is powered by synthetics, traders experience zero slippage on their trades. No, I didn't mean low slippage. I meant no slippage, because that is the power of the synthetics platform. No slippage on your trades. You can also easily short assets with iSynths, which are synthetic assets that move inversely to their target asset.

Synthetics isn't just for traders. Developers can build on synthetics to access the infinite liquidity offered by synthetic assets. Or investors can stake collateral to the protocol and earn fees that the protocol collects. If you're a trader and you're looking for a trading platform not found in the legacy world, check out quenta.io. If you're a developer or you just want to earn yield on your collateral, go to www.syntics.io, where you can stake your S&X or ETH and earn fees from synthetics.

Bankless Nation, we are pleased to bring you Vitalik Boutrin again in front of the nation. We are talking about a few things that are top of mind. Vitalik, great to have you. How are you doing? Thank you very much, David and Ryan. It's great to be on bankless again. It is fun. And I feel like this is a really good moment to be on bankless.

The first thing we want to talk about with you actually is this pretty momentous, maybe monumental in the history of Ethereum achievement, which is we hit the number. 524,288 ETH now deposited in the staking contract. How does that feel? It definitely feels great. I think for a long time, the first few weeks, and we were definitely a bit worried because, I know, the deposits were coming in slowly at the beginning. And it seemed like, oh, you know, with only four days left,

we only had 20%, and, you know, what the heck is going on here. And, I mean, we even saw, right, that there are a lot of people who were starting to get worried, like on our GitHub thread, and I saw Eric opened up that GitHub issue, basically suggesting, you know, hey, just in case there's never enough deposits, so it's set a maximum data, and there's a bunch of arguments around that. But, no, the Ethereum community came in and the Ethereum community delivered. And just within the last 24 hours, we ended up having more deposits than we had in the entire period before that. It definitely feels a little bit surprising, though in retrospect, it should not be too surprising.

I guess the one kind of events that happened in the past that this reminds me of is if you were here during the original ether sale back in 2014. If you remember that whole event, it was this kind of six-week-long sale, but then the first two weeks were this most important period because that was the period during which you could get the most favorable price for your ETH. And there was like some amount of deposits on the first day,

less on the second day, and then it seemed like it was going really slowly and almost petering out at a couple of millions. dollars and then just suddenly in the last couple of days everyone just like whooshed in and pushed the whole thing up to close to $18 million so and like the curve feels a kind of very similar to that and I guess to some extent you should expect that because well why would someone to participate earlier if they can just wait until the end and but you know just seeing that if spike you just come first to

and then all that ones is definitely amazing. And it's definitely a great testament to the Ethereum community as a level of confidence. That first time in Ethereum itself, and then the second time in the proof of stake and charting stuff that we've spent so many years working on. So, Vitalik, what does this tell you about the Ethereum community? Like, as a result of this,

because we hit the 524,000 ETH, but then we just blew past it by 200,000 more ETH. Like, what does this, if this is a litmus test for the community, what is it telling you? What are your big takeaways about the Ethereum community as a result of this event? I think, like, first of all, there's definitely wide community buying and confidence in ETH too in general. And that's something that I think I definitely believes the whole time,

but it's definitely good to have just such a clear, your sign of it. I know, $300 million worth of people just locking up their eth, potentially never to see it again unless the thing, you know, either delivers or refine or refine some kind of further thing to do to do to make those coins in the deposit is actually valuable again. So, you know, in some sense, it's the ultimate bet on progress.

And I think a bet on progress is to sub extent what Ethereum is about. I feel like it happened in the perfect way, like in a very, like, climactic way and really, like, fired the Ethereum community up when it actually happened in the way that it did. Like, it's almost fitting, I would say, that it, that we weren't sure we were going to cross the threshold until the last, you know, 48 to, you know, 24 hours. But I'm curious about this number itself. Why was the number selected? What kind of drove that number to begin with? was it fairly arbitrary that we have to have some amount of stake? And so, you know, why not this particular number? Or was there some sort of rhyme or reason behind it? I guess 524,000 was the,

of the power of two that seemed like a relatively minimum acceptable threshold for an amount of deposits that we could have to ensure that the chain would actually be safe with that amount. So like to give one example, it's the smallest power of two. that the Ethereum Foundation could not individually 51% attack. So that was, or the Ethereum Foundation has like slightly above that, but it would have to put everything in and even then it would only be around 50%. So if we had less, then just the number of actors that would have enough money to be able to take over the thing individually. would just get higher and higher, and so it actually would be, the chain would just not be that

secure with a lower amount. But on the other hand, like, at this level, it's getting to the point where, like, there's really very few individual actors that actually have the ability to, and have put in the funds to take over the chain. And aside from the foundation, there's also, like, obviously, the major exchanges, and maybe, like, consensus and a couple of hands, and I don't know how much they have. and a couple of whales. And then if you go higher and we hope to go higher,

then it is going to get out of reach of even those largest holders. But realistically, we don't have to get up to those much higher levels immediately. Like this is still, you know, just the phase zero beacon chain. There's nothing kind of completely relying on it. And so we start at a medium level of safety. and then go up to the higher level over time as people would naturally become more confidence in the system. So this process, this getting to that number

and exceeding the threshold of EF, that sort of initiates the rocket launch sequence, right? But the chain is not yet launched. That will happen December 1st, 12 o'clock UTC time. So what kind of happens next? What's the next milestone in the beacon chain launch and in staking now that we filled up the deposit contract to its minimum threshold? So the next milestone is obviously the thing that's happening in five and a half days after I'm saying this. December 1st, the launch itself.

We've done lots of test debts, but this is still larger than any test debt. It's a different set of stakeholders than any test debt. So just making sure that all goes smooth. Then after that, we would hope the proof of stake change just run smoothly for some amount of time. And at some point, we want to just start getting onto the task of kind of upgrading it and bringing on all the full functionality that we want ETH2 to have. So the big ticket items are, one is sharding. The second is none of the merge. So bringing ETH1 into the ETH2 system, removing the proof of the first.

removing the proof of work chain and kind of properly flooding the system back together. And then there's also some smaller things that we want to do. So, like, for example, we want to add light client support fairly quickly. So if you go into the EF2 specs at GitHub, there's already a PR that we've been working on for about a couple of weeks to add light client support. And then there is some just efficiency improvements and some possible kind of economic tweaks like one example of a type of economic tweak

we might want to make is just making the chain more friendly to people who validators who stay online during in an activity leak but don't stay online perfectly. So like basically if you have 90% of performance during an activity week

where the people who are completely online lose 40%, like you should not lose 4%. Really you should lose much less at 4%, is you really were doing the best that you can. So I have a proposal around fixing that just general efficiency improvements. So basically, the ETH2 chain is just going to kind of get into this, you know, progressing and upgrading mode. And it's basically preparing for the moment when the existing Ethereum ecosystem can

have properly and fully fold into it. Are you feeling a little bit Vitalik? Because so this is like the launch sequence initiated, right? But at some level, we're all kind of standing close to the launch pad watching the, you know, the SpaceX rocket. And it hasn't quite launched yet. And that, of course, is maybe the most nerve-wracking part, right? So are you, is there any element of like you being nervous about December 1st? Or, I mean, what could happen December 1st?

Could we have a failed launch? Could there be some issue? Issues are always possible, right? So this is, I think, why we should hold off the full celebration until it gets off the ground. But, you know, like we've seen what we've had with different tests. Some of them launched perfectly. Some of them had issues. Even the ones that had issues, they definitely all managed to fix themselves within either a couple of hours or a day or so. I expect that this time, there's definitely more

participants and potentially less experienced participants, though also maybe more experienced participants because the people who participated in the test debts are also going to be in the main net, and it's not going to be a first time for them anymore. We're not, I don't think we're going to see the thing that we saw in some of the testits where 40% of people just never showed up because, you know, this time we actually have a profit motive. There's just always the possibility of random technical risk, I suppose. And, you know, we've done lots of testing at this point. The quality of the test sets has definitely just gone up and up over time.

but, you know, figures crossed and, you know, it hasn't happened until what's happened. So we're definitely hoping and eagerly awaiting everything going well. So Vitalik, the ETH-2 research team recently did a Reddit AMA that I thought was really fantastic. And there's one question I recall in there. I think somebody asked you the question about maybe, you know, ETH issuance or something I forget the context for the question. You had this really, I thought, reflective reply where you talked about Ethereum being, you know, while EF2 is launching, it's also a technology and a network that is somewhat in motion, right? There are going to be a large set of improvements over the next two years, right?

It's not necessarily going to be like instant scalability is here and both chains merge. it's going to still be a process. But at the end of that time, like the ETH 2 chain will be in a much better place. I was wondering if you could, I'm not presenting it the way you did, the way you answer that question, but I'm wondering if that jogs your memory and you could talk about maybe that answer to the question like, what's the future for issuance? What's the future for stability of ETH2? Where are we going to be on the other side when all of this stuff is done? and why should we hang with Ethereum? Yeah, no, and I definitely think, I made that comment for a reason,

and I think there's an important kind of insight to keep in mind there. So there's a lot of people who have been recently commenting about how, oh, I guess Ms. Heath doesn't have a finalized issue in schedule, or, you know, you can't calculate what the current total supply is, or what is the max supply and now you're changing things. And there's definitely some extent to which those queries are coming more from the Bitcoin community than the Ethereum community itself. And so we should definitely kind of not, I guess, overrate the extent to which we

really feel the need to kind of respond and satisfy all of those requests. You know, like the thing that I think we need to just keep in mind is that just kind of being intellectually honest about the state of Ethereum, like, if you need a system which it satisfies the property that it will stay the same as it is today and it has existing properties that it's going to preserve, then Ethereum as it exists in 2020 is just as a matter of fact, not the system. you for at least the next one to one and a half years, right? There's just a reality that there has been this big, massive ETH 2.0 transition that's been planned pretty much almost since the project's launch. And now it's finally happening. And then at the same time, there are these massive transitions in economics. So EIP 1559 is a big one.

And also just activity moving from layer one to the layer two is another big one. And these things are going to fundamentally change how a lot of of economic and technical properties of Ethereum work. And so, like, any property of the Ethereum blockchain today may well change in those one to one and a half years. And so if you're interested in long-term economic properties of the Ethereum ecosystem, like, it matters less what are the properties of the system today. And it matters more, you know, what are the properties of the thing that's going to slowly

being built up over the next one and a half years, right? And like what even is the point of being able to, you know, verify the current, like, the current proof of work issuance schedule, like, perfectly if the thing that we really should be verifying is the economic property, some of the proof of stake schedule that are going to be, you know, guiding Ethereum's economics for the 20 years after 2020 or, you know, whenever. the merge ends up happening. So I think like just explicitly understanding that yes, Ethereum at the moment is a system influx and yes, if you want something that's not a system influx, then, you know,

Bitcoin may be the thing for you. Ethereum classic may be the thing for you, though. And Ethereum Classic itself is a kind of, I think, influx as a matter of necessity because it's a beatability is being threatened by all these 51% attacks. But so that's one example where, you know, in order to protect people's property rights. You actually have to do something other than standing still. So Vitalik, if this is successful, if we cross the bridge, if we make it for the other side to the other side, for those that kind of stick with the project,

stick like the builders who kind of stick with it, the community that sticks with it, what's left for us on the other side? What do we get in terms of issuance, in terms of improvements, in terms of security enhancements, What's the reward there? I think great things are on the other side.

So number one is obviously proof of stake. So much more efficient, much more secure, much less tree killing form of, much less issuance demanding form of consensus. So, you know, issuance is going to potentially go down by significant ability efficiency of the network will go up by a lot. And Ethereum is I basically going to become what I think the 2020s will expect a modern blockchain to be.

So that's number one. Number two is sharding. And at the same time, the layer two ecosystem. So those things together are going to give us scalability in the tens of thousands of transactions per second, which is hugely important because ultimately, you know, we're looking to, create a platform that can achieve mainstream adoption, create a platform that actually can provide a security interest that guarantees to its users and actually scale that up to millions and potentially tens or hundred millions of users and not just like a very small number, like a couple hundred thousand per day. And just proof of stake and sharding are the things that needs to be done

to make that happen. So that's number two. Number three is that just a better protocol economics. So EIP 5059 is the really big thing. I think the more you use Ethereum, the more you realize that just gas estimation is such a stupid and archaic concept. I really think that if all goes well in five years time, like blockchains that rely on this first price auction model

is just going to be looked at like caveman economics. So I'm really looking forward to it. It seems like the model is already working well and increasingly being validated in Fowcoin. It's increasingly working well on the test net. So really looking forward to that hitting the main net, I think that could be one of the big stories of 2021. Also, just general efficiencies within the Ethereum system.

So replacing the hex tree with binary trees, allowing stateless clients, which would make it much more efficient to process and to sync the chain, much more, some of virtual machine improvements that would allow us to just not need any more pre-compiles and be able to have all the advanced cryptography that we want inside of Ethereum. So basically just everything that you could want if you're, you know, building some like privacy preserving application or some kind of more complicated smart contract system or some or some roll up for scalability. You know, the tools actually are going to be there and you'll have and you're not going to have to use kind of clever tricks to anywhere close to the same extent as you do now. also light client support so in eith two we're taking a go at making it explicitly much more light client friendly and so um you know hopefully break the dependence on and if you're a and

kind of actually have uh you know make it so that people's wallets actually can be like clients uh so and fatalec do we get to that place not fixed issuance obviously like the bitcoin model but did we ever get to a place where we have more predictable issuance on the other side? What does that look like? Yeah. So I think realistically, we will want to wait around, you know, one or two years after, probably even the clock just starts ticking next week. So just after phase zero and just verifying that the current economic structure of proof of stake is sustainable,

verifying that that economic structure can survive the addition of. sharding the addition of the EF1 chain verifying that EIPO 1559 is working as expected. And then once all of those pieces are in place, then I think people are going to just have a much more clear and visceral understanding of, you know, this is how high proof of stake issuance is, this is how high the AP 59 burn is, this is the kind of historical variance of those two parameters. And so the issuance in general kind of change of the total supply level is going to look, I think, much more predictable a couple of years from now. And I think people are going to be very happy with the results. Last thing for you on this, Vitalik, then we'll move on to something

that is equally exciting in a way. If somebody is just built on the ETH-1 chain right now, has all of their ether on the eth one chain. Do they have to worry about any of the stuff that we were just talking about? Or does this merge somehow in the future? Good question. The answer is they absolutely do not. So one of the big compromises that we made for kind of practicality is that while Ethereum's consensus and scalability and all of these things

are getting massive kind of revamps from the ground up, Ethereum's execution layer is actually remaining remarkably constant throughout the transition. So transactions that were valid before will be valid after, smart contracts that executed in one way before, we'll execute the same way after. Basically, all applications, contracts, eth, everything will just be kind of migrated from the ETH one side to the ETH2 side automatically. And so you as a user or as an application developer basically do not need to worry. I mean, there are a couple of corner cases.

Like, for example, if you relied on block hashes for randomness before, then in proof of state, block has stopped being secure randomness. And so instead, you would have to use, like, the RANDA output as randomness. But, like, these are, block times are going to work slightly differently. But, you know, these are generally, like, issues that only affect a very tiny portion of users. So for basically, like, the fast, vast majority of the user base, I think they'll experience. and of close to no change from an application perspective. So, Vitalik, the conversation about proof of work versus proof of stake was really, really big

from what I can remember in 2017 and 2018, or it was probably even bigger before that, but I came into the world of crypto in 2017. And I actually think it's really valuable that people understand this debate and why, and understand the merits of proof of work and also understand the merits of proof of stake in order to understand, like, why is Ethereum so committed to, proof of sake. Like, why do we want this? And so you wrote this article on your blog, Vitalik.ca, everyone to check it out, called Why, Proof of Steak? And we kind of want to go through some of these points so we can help the bankless nation, like kind of come to understand the terms

that you use here and the arguments of why proof of stake is inherently the right, I'm not going to say superior, but the right consensus mechanism for Ethereum, right? Because many Bitcoiners will still say to the stay that proof of work is the one true consensus mechanism. So let's go through some of the points that you that you wrote here. Proof of stake offers more security for the same cost. Like what does that mean and how does staking offer that? So the basic idea here is that if you try to quantify security, so the easiest way to quantify security is just how many dollars do you need to spend to break the thing. Then if you look at the ratio of security to cost where cost just means like how many rewards you have to push out to the

participants, then the ratio of the security level to the two of the ECAM amount you pay for the security just is much more favorable on the proof of stake side than it is for any version of proof of work. And I think the way that you can look at this, right, is that if you look at, just first of all, start off with a comparison between GPU-based proof of work and ASIC-based proof of work. So this is a comparison. that even Bitcoin people make all the time, right, when they defend why their ASIC-based proof of work is better than the ESIC-resistant algorithms. It's basically that in the GPU case, the cost of mining is only operating costs, and there's basically no capital cost, right,

because, like, you can rent GPUs. If you buy a GPU, you can always sell it after, even if the blockchain disappears, GPUs continue to be useful. And so if you want to to attack a GPU-based chain, all you have to do is just rent some GPUs for six, hours or however long you want to do an attack. And so the cost of an attack is basically going to be just exactly the same as the cost of the rewards during the time period that you're attacking. And if you as an attacker receive block rewards, then the cost of the attack even drops to zero or it becomes profitable all by itself. Right. So for GPUs, like, while GPU proof of work mining has a big advantage in terms of being decentralized, from a security perspective, it actually is

quite weak, and we've seen this, right? As I mentioned, in an Ethereum classic, has been getting a lot of 51% attacks recently. Now, if you look at ASIC-based proof of work, ASIC-based proof of work is considerably more secure because ASIC-based proof of work has not only ongoing costs, but also capital costs, right? And an analysis I did a few years ago suggest that it's about one-third ongoing cost and about two-thirds capital costs. Maybe that ratio changes over time. I actually have no idea which direction it's going to change into, but like that's roughly what it is approximately, right? And if you, then if you kind of make some assumptions about how long in ASIC lasts, you can actually kind of reverse engineer what the capital costs are,

kind of from the amount of money that you're paying for rewards, right? So the idea basically is that if your chain is paying $1 a day in rewards, then out of that $1.3rd will be operating costs, and sold be capital costs. And you can think of the two-thirds that's capital costs as being a kind of amortized over the entire period of time during which that miner is mining, right? And so if you assume that an ASIC lost about two years, I think now it's a little bit more, it's a little bit more than two years. But, you know, Morsal increasing efficiency, wear and tear all of these things.

So at the two-year level, right, all you have to do is you basically take the amounts per day that they're spending on the ASIC, and then multiply that by the number of days in two years, and that's the cost of the ASIC, right? And so if your model of an attack is that an attacker just has to buy up as many ASICs as a legitimate network, then you can use this to kind of get an estimate for the cost of the ASICs needed to attack the network. And so here, you get about Fortune 86 dollars, and so that's the cost of breaking proof of work if you have ASICs, right?

So already much better than GPUs, but ASICs have a cost. cost of centralization. So now if we go to proof of stake. So before we say, can I get you to comment on Ethereum 1? Would you, if Ethereum 1, if that was the chain that would just carry on into the future forever, would you have been in favor of it being GPU based? Or would you have been interested in making an ASIC version of Ethereum? So at the beginning, I was definitely much more pro-GPUs.

But I think more recently, well, it's not so much that I've become more pro-ASIC, it's that I've become less pro-GPU. And so it's just like, it just feels like kind of a tragic choice between two differently bad options, to be honest. But GPUs did have this nice property in the early days that it was a much more egalitarian, decentralized way to distribute the tokens, which arguably, that's kind of what Ethereum was looking for in like the first part of its life. Yeah, I agree.

And I think that's very important. And I think Ethereum is definitely better as a result of having that GPU mining for the first few years of its existence. As somebody who came into Ethereum because it was able to be GPU mined, I have to say, thank you for making that choice. But not as secure compared to A6 is the point, at least from an economic security perspective. And the thing to keep in mind is that the longer we, I think the chain even manages to be a GPU mind up until this day only because people realize that proof of stake is coming, right? Like, basically the reason why there isn't much more work on building A6 is that people realize that any ASEX that get built would only survive like either a year or potentially even a few months, like potentially in response to the chain getting taken over by A6, we could even just roughly.

the entire transition and just like do the merge in a couple of weeks, right? Like doing that sort of thing would totally be extremely disruptive to the Ethereum ecosystem, but you know, doing something extremely disruptive in response to an already extremely disruptive attack is the sort of thing that would happen, right? And so like people, I think, realize that it's just too risky to build A6 because like there's just no guarantee that you're actually going to have enough of a period of time during which you're going to be profitable, and that's why the chain continues to be GPU-friendly, right?

And so the facts that, like, this GPU-friendlyness has been possible to this day should not at all be taken as evidence that GPU-friendliness is a stable equilibrium going into the long-term. Right, makes sense. Shall we get into proof of stake? Yeah, so proof of stake.

So basically, the way that I view proof of stake is I view proof of stake as combining the efficiencies of, or this capital cost property of ASICs, but really cranking that property up to the extreme and at the same time having the kind of open accessibility of GPUs. Right. So basically proof of stake is almost entirely capital costs, right?

The cost of like operating a node is fairly small. And the difference is though is that in the case of ASICs, the capital cost is only kind of amortized over the course of, you know, about two years. So it's about like 50% a year. But unlike A6, deposited coins do not depreciate deposited coins. You can deposit them. You can stake for a bit and you can get them back.

And so you actually don't even need that high if a rate of return to motivate people to stake. So here I was actually, I think, understating the case, right? Because here is saying, oh, you know, imagine a 15% rate of return is enough to motivate people to be willing to walk up their money and stake. And I think 15% is roughly the rate of return. return that we're going to see in this initial phase of youth two. Like I think with the current level of deposits, it's about 18%. But, you know, we're going to see more deposits. But I think in the long run, especially when it's possible to just deposit, um, stake for a bit and then we'll draw, get your money back within a day. Like, I fully expect that the rate of the, the rate of return

will just drop to something closer to 10% or even 5%. And so, but even with 15% rate, a 15% rate of return on capital, that basically means that, you know, $1 per day of rewards is going to attract $6.667 worth of deposits, right? Or basically, like, if you have a two, if you stick $2,400 in asset deposit, then your annual return would be $2,400 multiplied by 15%, which would be around, you know, close to $365. And so $365 a year is equal to $1,000. to $1 a day, right? So if we assume 15% rate of return per year, then every $1 of day in rewards is going to correspond

to $2,4003 of deposits. That's at current rates, if the rates drop to 5%, and this number obviously goes up to $7,500, right? Now, if we assume that only 90% of the costs are capital costs and there's still 10% that are operating costs, then the 2433 goes down to about 2189. but basically even still, right, we have five times more deposits that are paid for by the same

rewards in proof of stake as compared to ASIC-driven proof of work. And this is now in the long term, like I expected to go up to something like 10,000, so it will be 25 times stronger than the proof of work. So does the security, so is that summarized as the security of proof of stake comes from the fact that ether, the asset, doesn't deprecate in the same way that ASIC a machine does. So ether gives you a year-on-year return, where an ASIC gives you a year-on-year deprecation, devaluation of the unit, right? And so that's the fundamental component of where proof-of-state gets its security from.

Yes, that's definitely an excellent way of stating the point. Cool. All right. So where we've come so far is we have GPU-based proof-of-work, which Ethereum is on now. That means you can use a GPU card and like your Nvidia, like, gaming, machine to mine a little bit. And that total cost of attack is like fairly cheap. You can attack that fairly easily, lower economic security. And then Bitcoin has ASIC-based proof of work, which is a bit more expensive to attack. So higher economic security. And that economic security comes from the

fact that that ASIC is only doing one job, that's only possible to do one job, which is mine Bitcoin. It's dedicated. So it's all sunk capital costs, basically. You are locked in. So, and now proof of stake, because it's like ultra sunk cost, that's the only thing you can really like kind of, you know, do with it. It's almost like a tokenized ASIC on steroids, if you will. That's even more efficient.

And how did you quantify this? Is this like a 5x more efficient, like bank for the, bank for the security for the dollar? Or is it a 10x? Like according to the kind of back on my back of the envelope numbers here, it's 5x in the short term and up to 20x in the long term. Wow. So I think of this, you know, Dave and I have, you know, often sort of analogize these sort of chains to nation states. And almost this is like the security force. This is the defense, right?

And so it's almost like you're upgrading your military from cavalry and horses to like pans or tanks here, right? Like trying to think like what's a good a good analogy like okay so here's a really fun and dumb analogy So GPU based proof of work is you have to like basically hire soldiers out every every day to patrol your borders To make sure the enemy army doesn't get in and you have to keep on paying them Basic based proof of work is the great the great wall of China Then proof of stake is imagine a great a Great Wall of China that has the property that the way that this wall is constructed is that when people are not staying in their own houses, they just teleport the walls of their

walls of their houses and they make it part of the, they make it part of the Great Wall of time. So whenever you're not in your house, whenever you're like going off on vacation, you just click a button and your house teleports to being part of the wall. And, you know, the government pays you some small amount of money to do this and you're happy to collect the reward. And then as soon as you want to move back into your house, you just click a button again and the balls teleport back at their part of your house again. So basically, it just, like you don't even need to kind of create these separate resources. You're just kind of taking advantage of existing resources and you're moving them over to a different purpose when people know that they don't need them for a long time. So much more capital efficient.

So the idea is that the Great Wall, this version of the Great Wall of China would theoretically be a lot higher because of the, incentive for everyone to allocate the walls of their house into the walls of the Great Wall of China in the moments that they are not inside their house using their house as a house. Exactly. The interesting thing, so is I think here, you talked about this in the article, right? So we're talking about a 5 to 20x gain in proof of stake potentially. And the question is, what do we do with this efficiency gain as a social contract for Ethereum. Can you talk about that? Because it seems like there are like, you know,

maybe two options here of what to do with this drastically improved military, drastically improved great wall. Yes. So one of them is to keep the military budget the same, but have much higher security. And the other one is to have the same security, but cut the military budget down. And I guess the reason why I think we're taking the second approach is actually best I've got to even described to the next section, right? I think Ethereum has enough security today, especially if attacks become much easier to recover from. And we actually don't even need to have too much spending going into keeping the wall up because, you know, it turns out that basically even if someone successfully invaded. so like you can do this clever move, clever move where you just like hard fork and you delete their army. So.

So I want to get into that. So just to link folks into what you're talking about. Like so the way Ethereum right now and the way Bitcoin right now pays for their military is generally through block reward, through issuance, through supply, right? So and Bitcoin's approach is to cut that military budget, at least military budget denominated in Bitcoin every four years, cut it in half every four years. So we're the Bitcoin protocol government every four years. You know what? It's forced on you.

We're going to slash your military like by four, right? At least denominated in Bitcoin. By two, by two, excuse me. Ethereum's approach is basically to do something that's maybe more, what, the social contract is kind of called minimum necessary issuance or minimum viable issuance, where it's like, hey, like, maybe we can drop this a whole lot more because we have a stronger defense system. And so that that issuance, essentially, is what goes to pay for the military. And maybe that could be slashed. Maybe that could be cut. Right now it's a four, four and a half

percent or so. And we could go to something less than one percent. Absolutely. Absolutely. One of the theories I've had about one of the reasons why proof of stake is more secure in the long term is that with that reduced issuance, that reduction of issuance actually compounds as a benefit over time. And I want to get your take on this because I don't know if this is real or not. And the same dynamic is present in EIP 1559. The more ether that's burned in EIP-1559, the more ether is unable to be leveraged in an attack against Dakers, right? The less ether we can make available to the secondary market, the less ether there's available for attackers to be able to grab and use that to attack the Ethereum network.

Is that a real thing? So like the less ether that we issue, is that over time adding to security? that's definitely an interesting and a kind of novel way to think about it and I guess and generally yes like the less ether there is and and even like the higher the price goes so the harder it is for an attacker to buy enough to successfully make an attack and that's definitely one of the factors that contributes to security and if you have miners that are just constantly selling millions of coins every year. year, then, you know, some of those coins are going to end up in the hands of people that might be interested in attacking it at some point. So let's go into one of the grand ways that say somebody did somehow manage to get their hands on enough ether to attack the chain. We actually have a way around this. And so one of the sections in this piece is titled, Attacks are much easier to recover from in proof of stake.

Maybe we can talk about after an attack happens, what paths to maintaining or, you know, restoring security does Ethereum have? Yeah. So I think in the GPU case, the only defense is basically just waiting and hoping the attacker goes away, which has worked so far, right? But as soon as an attacker comes as persistent, you're basically screwed and you have to try to figure out something else. in an ASIC based system, you do have one form of defense, which has you just change the proof of work algorithm. And what that does is it basically kind of bricks all the ASICs that renders them useless. But it bricks both the attacker's ASICs and the honest miners ASICs, right? But if the attacker is willing to suffer that initial expense, then at that point it just turns into a GPU mine's coin up until, you know, after a couple of years where people build ASICs for the new algorithm again.

And so it just turns into the first paragraph and the attacker can just continue the attack really cheaply for as long as they want. In fact, I would even argue that it's worse than this because if an attacker starts doing this and if the attacker commits to continuing to attack, then it actually takes away from the incentive to build A6 for the new algorithm because anyone building the A6 knows that if the attacker is going to continue, then potentially those A6 could get brick the second time. So basically the fact that breaking hurts both the attacker and honest miners actually it really makes the game theory behind all of this even worse. So in the Proof-State case, basically it gets much better, right? Because we have this concept of swashing, right?

So if you contribute to breaking the chain, you get slashed. And even if you contribute to breaking the chain as part of a 51% attack, you get slashed. Now, there are attacks that are hard to detect and hard to slash forth. So a big one is like a 51% coalition censoring everyone else. But you can get around this by using this wonderful trick that the Bitcoin people invented in 2017 called the UASF, the user-activated soft fork, right? Basically, just users coordinate on participating in a minority chain and they just abandon the majority chain that's censoring them. And actually because of the way that the Casper FFG rules work, the attacker cannot legally kind of move over to the minority chain without contradicting themselves and getting slashed. And so the attacker just gets weak and they lose about half their points, right?

So basically no matter how you attack the chain, it's going to be very expensive for the attacker, not very expensive for everyone who is not an attacker. And so the game is very asymmetric, right? Like every time the attacker attacks, they lose a huge amount of money. and potentially like the the amount of coins that are burned as a result of an attack could even mean that an attack would cause the price of ETH to go up. So like one of my favorite features about proof of stake is that if an attacker attacks Ethereum, it makes all the stakers rich by default. Well, not by default, but it's highly likely that the price of ether goes up because

of somebody's buying up all this ether. One question I had when I had about this. like using this as a defensive mechanism in proof of stake, this is relying on the social layer, right? This is in this moment, security has migrated away from the actual code and moved into the social layer, which maybe it perhaps is one day, if this ever does happen, it would be coordinated via something like Twitter or Reddit. Can you talk about that dynamic? Yeah. So there's definitely a lot of discomfort in relying on the social layer for consensus.

But actually, I wrote about this quite a bit in my original 2014 article, the one where I'd say, called how I learns to love weak subjectivity. If you actually even just go in and Google for that now, you can probably find it. So basically, the arguments that I make is that relying on social consensus is actually reasonable occasionally. It's just that social consensus is not something. that you want to keep on it keep leaning on forever right so the idea is basically that like i mean even bitcoin relies on social consensus to some extent right like even you know the uasf that was proposed in uh 2016 was ultimately a form of social consensus overriding mining so

basically um you know hard forks are done with uh social consensus um even the idea that if miners push a blockchain that would increase the supply to beyond 21 million, that people would reject that blockchain is ultimately secured by social consensus. The difference between a malicious protocol change and a bug fix is social consensus. So relying on social consensus on these kind of extreme matters, and especially on matters where there's clearly one side which is an attack and one side which is just trying to push the chain forward is, I think, actually, a reasonable thing to do. The thing that you don't want to do is you don't want to rely on social consensus too often, right? And the thing to keep in mind is that if you even think about, like, how an attack on proof of work would be responded to, right?

Like even like defenders of a basic base proof of work will say, oh, if our 51% attack happens, we can just break the ASIC and we can change the algorithm, right? Like if you press them on it, they will say that this is a defense. But that defense as well, it's also a social consensus. Like changing the per work algorithm is also a form of social consensus overriding the overriding the code. So it's also striking that the price of Bitcoin, the price in itself is one layer of social consensus too. It is indeed. Yeah.

And we've seen we've seen this, the idea of bricking A6, that's played out a number of times, I believe, in the Monero community. So this is not hypothetical conversations. This is strategy that blockchain ecosystems have leveraged. AVE is a borrowing and lending protocol on Ethereum and just recently released Avey version 2, which has a ton of cool new features that makes using Avey even more powerful. With Avey, you can leverage the full power of defy money Legos, yield, and composability the all in one application. On AVE, there are a ton of assets that you can deposit in order to gain yield, and all of those same assets can also be borrowed from the protocol if you have deposited collateral.

Here you can see me getting a 200 USDC loan against my portfolio of a number of different defy tokens and ETH. I'll choose a variable interest rate because it's a lower rate than the stable interest rate option, but I could choose the stable interest rate option if I wanted to lock that interest rate in permanently. One of AVE's V2 features is the ability to swap collateral without having to withdraw your assets, trade them on uniswap, and then deposit them back into AVE. AVE does all of this for you, all in one seamless transaction, so you don't have to repay loans in order to change the collateral you have backing them. Check out the power of AVE at AVE.com. That's AAVE.com.

If you want to live a bankless life, you need to get a monolith defy visa card. is a one-to punch of both an Ethereum smart contract wallet and an accompanying Visa card that lets you spend the money that you have in your Ethereum wallet everywhere where Visa is accepted. When you swipe your Monolith Visa card at the grocery store or at a restaurant, it actually makes a transaction on the Ethereum blockchain that spends some of the money you hold in your Monolith wallet. It's insanely cool and it's one of the best tools out there for living a bankless but still normal life. Monolith also offers on-ramp services for getting your fiat money into the world of defy. So it's trivial to top up your monolith card if you ever

need to and your deposited money goes straight into your non-custodial wallet so your money is never held by a centralized intermediary. Because monolith is native Ethereum infrastructure, the money you hold in your monolith wallet still has the power of defy behind it. Swapping assets on uniswap or earning yield in defy is at your fingertips. Go to monolith.xyz and sign up to get your monolith and beats a card today. So let's move on to one of my favorite subjects, which is how proof of stake is more decentralized than A6. Some people would call this a bold claim.

Vitalik, why can you make such a claim? So, and I think we do need to break up proof of work into two categories, GPUs and A6. So, you know, like there is a very decent case in GPU based that proof of work is quite decentralized, and I'm happy to admit this. like my quam against the GPU mining is not about the decentralization. It's about the security. Now, the problem with ASICs, on the other hand, is that in order to make an ASIC, you need millions of dollars of capital. Right.

Now, you could say, oh, you know, you don't have to make your own ASIC. You could buy an ASIC. But the reality is that if you're buying an ASIC, the side that's making the most profit is not going to be you. It's going to be the ESIC manufacturer, right? People selling the shovels. Exactly. Like in the long run, it's like the game is very much one that's kind of skewed to benefit whoever can sell the best shovel. And, you know, whoever can sell the best shovel has a lot of, like, they can potentially even like sneak in code that could allow that ASIC to be bricked remotely.

Like they could do all sorts of real evil things that are harder to do with if you're running the thing on general purpose hardware. So basically, like, ASIC-based mining really does rely very heavily on this very small number of these very wealthy and powerful manufacturers. And so, like, basically, if proof of stake means the rich gets richer, well, guess what? AIC mining also means the rich get richer, except the bar that you need to pass in order to get richer. In the case of ASIC mining is something like $10 million. But in the case of proof of stake, it's $20,000 today. much less than $20,000 in the future when staking pools become possible. So one of the, I've been chewing around the ideas of staking pools,

and there's a bunch of different design parameters. But I think the staking pool that retains as much ethos of Ethereum as possible, which I think there is a spectrum of how much does a protocol or company on Ethereum retain the original ethos of the system. There's something to be said for that. And that's kind of where I turn my attention to Rocket Pool, where Rocket Pool is trying to become a staking pool for people to decentralize the amount of people that can participate in staking by being a pool, and then also decentralizing the nodes by making more total nodes available to be part of the Ethereum Proof-of-State Consensus Mechanism.

When you think of Rocket Pool, is that what you think of? Yeah, I think Rocket Pool is definitely one of the good examples. examples of someone trying to do a good job. The ETH2 protocol is really explicitly designed to make a stake decentralized staking with things like multi-party computation as easy as possible. So that's definitely the sort of direction that we're looking into. There's also this sub-argument that you made here, Vitalik, that proof of stake is more censorship resistant.

And that's basically because it's not on the electricity grid, right? for instance, like in the way that a GPU mining or ASIC mining footprint might be. Is that the main argument here? Yes, exactly. Like, if you imagine, you know, you're living in, say, Venezuela and you want to participate in the Ethereum network, then if you're a proof of work miner, then, you know, realistically, you're consuming huge amounts of electricity, you have to get huge amounts of hardware. If the government wants to find you, they can find you, right?

It's not hard to say someone consuming that much electricity. These are techniques that they've already perfected for years to figuring out, finding people who are growing the wrong kind of plants in their backyard. So, you know, this is fairly established. It does make you wonder how long it would take a motivated government body like the United States or some other government body. Maybe let's not name different governments here to go see. out all of the major mining, Bitcoin mining, or even Ethereum mining facilities in the world and

just confiscate them, take them over by physical force. These are off now. Yes. It's just, I think it's kind of a little bit of a game maybe that summon crypto play, as if we've already faced off against these nation states that are actually trying to censor us. The reality is they're just not even trying yet. So, yeah, that is an advantage that may come into play.

But there are also some advantages that you talked about here, Vitalik, of proof of work. And I think we'd be remiss if we didn't mention the way proof of work actually is superior because there are some tradeouts. Can you talk about these? Yeah. So these are what I think are the two best arguments in favor of proof of work. Like one is this critique that proof of stake feels more like a closed system. like there isn't this kind of new source of issuance that kind of brings fresh economic blood into the ecosystem.

And so, and it feels like the system has this kind of naturally, kind of intropic pressure to just be more of the same people by default. And this is definitely something that does kind of concern me to some extent. And the best rebuttal that I have is that in proof of stake, the rewards in general are going to be fairly low, right? Like, I think if proof of stake rewards were, say, 10% of, like, annually, then, like, I could see that very quickly lead to a few stakers very quickly. I'm having a very large portion of the entire supply and just dominating everything.

But here we're talking about validator rewards equal to, like, up to 0.5 to 2%, probably close to the 0.5% of the total leaf supply. And the more validators are staking, the lower the interest rates get. And so the amount of time needed for this factor to cause a large amount of concentration is huge, right? It's potentially over a century. And on such time scales, like, doesn't really matter much, right? Like, people want to spend money. They want to distribute their money among their children.

They want to give their money to charity. And so, like, aren't the, like, basically, if there are any natural economic pressures toward deconcentration at all, then, like, they're going to have an effect that's larger than 0.5% every year is basically the argument. Whereas if that rate was instead of 0.5%, something like 4.5%, then I think this argument would actually have a much stronger case. So this is one of the reasons why I'm pushing for lower rewards. And the reason people make this argument is, of course, is as we stated earlier, steak d'ath is almost like an everlasting asic. There's no depreciation.

There's no capital depreciation, unlike a real world. and it doesn't deprecate. Exactly. And notably, the people that crypto has, like, accidentally and inadvertently made really wealthy are largely people that, I'm guessing, this is my gut take, that weren't wealthy beforehand, right? And so, like, what's the point of hanging on to you every single last denomination of ether if, like, you can't sell it and, like, have, like, kind of a nice, more comfortable

lifestyle? Like, the people that have large supplies of ether are going to have to sell. if they ever want that material impact to do something real in their lives. And so that's kind of what I think what you were alluding to, Vitalik, is that the rewards, the passive income rewards to ether accrue slower than people's, people find ways to sell their ether to buy a new house or new car or something. And so there's like this natural sort of incentive to let go of your ether. Well, what about the second point, Vitalik?

We talked about weak subjectivity earlier. Does anything bear repeating here? No, weak subjectivity is this extremely fascinating topic. And I think it's been really the center of the debate almost ever since the beginning and up until this day. Like basically, weak subjectivity is this really subtle thing. The idea is that in proof of work, you have this property that you have your client and you can just start it at any time.

even if you've been offline for 10 years, you get just started up and you can seek to the valid chain and you have all, or the canonical chain and you have all of these that economic guarantees. In proof of stake, like you can still do that, but it depends on what guarantees you want.

Right. So like if you're guarantee, if the guarantees that you care about are in an honest majority model, so you just assume that the majority of people are nice, then you can still do that. But if you want to have a hard economic guarantee, so you want to have a guarantee,

that says, you know, tricking you requires the attacker to burn a huge amount of money, the way that proof of stake kind of promises a strong economic guarantees, then you lose this property to some extent, right? And the reason why you lose this property to some extent is basically because, like, if an attacker just create, like, who has, say, 51% of the stake creates a fake chain and they convince you that this fake chain is the real chain, then, like, normally the way that they will get punished for this is it would be slashed, right? Like, if you create a fake chain that conflicts with the real chain that you already contributed to making,

then, you know, some large portion of the validated ZootenSACs, you can make proofs, and you can publish those proofs into both chains, and you can burn the attacker's money. So, right, this is just, like, standard, you know, proof of stake theory. like how swashing works. But the problem is, what if the attacker has already withdrawn their deposits that they use to make the attack by the time that they actually need to do the attack, right? So what if someone has been offline for a year and you feed them a chain that looks like a valid chain from the perspective of their client that was online for the last time

a year ago, but back in present-date reality, you've already taken out all your points, and so whatever slashing happens can't actually hit you. So you could still trick them, and you would not get slashed for it. And so if you want to be resistant against this attack, basically, you just have to log on and kind of be online often enough that basically the time difference between any two times during which you come online is less than the time that it takes for something like one third of the validator set to be replaced, right? Which is, and that number depends on the size of the validator set.

It's like somewhere between a couple of days and a couple of months, depending on how many people are participating. So like that requirement, the requirement that in order to have a really economically guaranteed level of security on which chain you're getting, you need to log on with your client at least once every few months. Like that is the this week's subjectivity requirements. And the reason why

a lot of people feel uncomfortable about this is because it basically means that if you're just joining the system for the first time, then, you know, you, for the first time that you log on, you need to basically find some third party source

that basically tells you like which chain is the canonical one that you should be following. Now, most of the time, this actually is actually isn't a big deal because most of the time there aren't going to be 51% attackers trying to use their old keys to screw people over and so there is really going to be like one chain that has enough signatures right but in the extreme case where there are attackers that can like buy up people's old keys then basically in order to disambiguate you do need to find someone to trust to give you that information and I think

fans of proof of work say that you know oh my god this is trust this is fatal what what are you doing? You've completely destroyed the point of cryptocurrency. But my argument is basically that like this level of trust is a level of trust that we already have. So like for example, you already trust developers to tell you the most recent hard fork version. You know, you already trust them to fix bugs. You already get soft corrupted from them once every, you know, half a year. And so why not just like leverage that existing level of trust? And I use it for proof of stake checkpointing as well.

So that's basically the... So with proof of work in Nakamoto Consensus, we have the ability to actually legitimately and concretely verify the amount of work that is done on a particular chain. We don't have that with proof of stake

because we're not doing work. We are, and so there... We're verifying signatures that are signed by validators and in order to recognize that those validators actually are real, like you have to look at the state of the blockchain itself, like at time, you know, a couple of months ago to a year ago, right? So the security of approved stake chain is definitely more kind of, it's not like, it's not circular.

It's more like a spiral, right? It's like the security of the system today depends on the validity of the chain six months ago. And the security of the chain six months ago depends on the validity of what happened 12 months ago. And so like if you skip one of those steps, then like, and if you're on offline, for longer than that period of time, then you do need some way to kind of help you disambiguate and come back online. But if you're there more often than,

like basically staying in sync with the chain is more of this necessarily dynamic process. Right. Right. So here's kind of, I guess, my takeaway to just sort of summarize this article. And you guys jump in if you have some other things. If you are willing to get past this edge case, weak subjectivity requirement. And you made some compelling arguments why that that makes sense,

or it's less important than others think. And if you're willing to get past this issue that stake never depreciates, and so there is the capability for people in the system to continue accruing wealth. And you sort of mitigate against that, against that with lower rewards overall. Then what you end up with is a chain that is much more economically secure. And also more efficient in that economic security. So you can do things like reduced issuance, reduce the cost to actually secure that network. Essentially, you get- More security per dollar.

Exactly. That's what proof of stake. That's what proof of stake provides, at least under Ethereum, right? Because there are other flavors of proof of steak. But would you add anything to that, Vitalik? No, I think in general, that's a great summary. Very cool. And Vitalik, when a lot of the design design design design,

decisions behind proof of stake. And when you, when you, you know, verbalize your arguments and kind of just like thought out loud about these, the weighing the pros and cons. The way that you did that turns to our, are the next article that we want to talk about, which is convexity versus concavity. And so, like, and we're just burning through your articles here. So you can put out a blog post faster than anyone. And I totally aligned with this article. And Ryan and I chewed on it together and had some conversations. about the implications of this. Before we get into those conversations,

maybe you can kind of give us the synopsis on, like, convex versus concave dispositions. Yeah. So I think the core idea here is basically that, like, whenever there's a tradeoff and there's, like, a choice between kind of taking one path and taking another path, one of the kind of more subtle differences between how people view that tradeoff is basically, like, do they, is how they view the kind of extremes as opposed to what, like, the options in the middle, right? So basically, do they view it as the sort of situation where a moderate option is kind of more likely to be better than the extreme options, or do they view it as this scenario where, you know, it is one or the other? And if you try to mix the two together, then it just completely doesn't make any sense. And, like, you know, it really is either this or that.

And so like those two graphs that are right there, like they do a good job of kind of showing what the perspective is, right? Like basically like the graph on the left, like the concave worldview is like kind of like you have a graph that's kind of sloped downward. So it's kind of like a hill, right? Like so there's there's always some option in the middle that's best. But, you know, whether it's kind of in the middle but much closer to the left or on the middle, much closer to the right, that's something you can legitimately argue on. But then, you know, as you go further to the extremes, then eventually on either side, it just starts becoming worse versus the convex worldview, which looks less, the chart looks less like a hill and more like a valley where, you know, you pretty much, like you want to be either on one side or the other side in the middle is the worst place to be. So, again, like, this starts off being very abstract, but if you go, like, once we kind of go forward and go a few paragraphs down, like we can see it where it starts becoming more concrete and where I can see some like specific examples of like what concave and convex means.

Right. So here I just like give us some examples of like what kinds of things someone with a concave disposition versus a convex disposition might say. So like one of them. So like one kind of concave statement going to the extremes has never been good for us. You can die from being too hot or too or too cold. we need to find the balance between the two that's just right. If you implement it's only a little bit of a philosophy, you can pick the parts that are the best and avoid the parts that are risky, but if you insist on going to the extremes, then once you've picked the low-hanging fruit,

the next fruit that you have to pick are going to be forced to look harder and harder, and you'll have smaller and smaller benefits, and the risks might multiply and outweigh the whole thing. The opposing philosophy, the philosophy that you favor less, probably has some value too, and so you should generally try to combine the benefit, the best parts of both, and definitely avoid things of the opposing philosophy continues to be extremely terrible, just in case they have more of a point than you realized.

And then some examples of a convex statements. One is we need to focus, otherwise we risk becoming a jack-of-all-trades master of none. Another example of a convex concept is the kind of the slippery slope concept. So if we take even a few steps down that road, it will become a slopeary slope and only pull us down even further until we end up in the abyss. There's only two stable positions either here or there, and we want to stay here. If you give an inch, it'll take a mile. Whether we're following this philosophy or that philosophy, we should just be following something and stick to it. Stick to it.

Making a wishy-washy, a mix of everything doesn't make sense. And so after this, like I get into what's probably, the kind of motivating example of this idea for me, which is basically the Bitcoin versus Ethereum. Well, like, before before before before we get there, like I feel like I've met these two types of people. Right. Like I've met the, you know, the binary thinker, the convex disposition thinker where it's just like you're with me or against me kind of individual, right? And then I've definitely met the the concave disposition.

which is a bit more the pragmatist. I could see it this way. I could see it that way. Why don't we find the balance in between? And one of those definitely sounds like more of an Ethereum, not to stereotype. And one of those sounds definitely more like a Bitcoin or, again, not to stereotype. Part of me wonders how much people are just hardwired into these modes, like, from like birth and genetics and, like, whatever happens to them in the early ages. because I don't know.

I've definitely met these types of people and I'm not sure how much they can kind of help it. And it's definitely noted in a crypto industry. Like when you come into this world, do you find yourself quickly moving into a camp of sorts? Like usually bitcoiners or Ethereum's. And like you kind of find yourself in to be one or the other. And there never really seems to be any of this like moment of flux

where people are confused. They just kind of come to find themselves aligned with one of their particular communities. And I think you've totally tapped onto a very real parameter of personality that I think would predispose someone for one community or the other. No, definitely, I think. Again, it's definitely a really interesting question. Like, to what extent, like, someone's disposition toward this, like, is, you know, either genetic or just, like, some cultural thing that's, like, so deep, like, 5,000-year-old or whatever that I might as well be genetic.

Like, if that's true, then, you know, that does mean that, you know, for example, even if you personally like Ethereum more than Bitcoin, that there's just no way that you're ever going and convinced 100% of the world population to also be in that same camp. And so, like, you just have to acknowledge that Bitcoin is going to continue to exist and there will always be some constituency that it appeals to. And if you're a Bitcoin person, then you'll have to kind of pretty much admit the opposite. So, you know, if people really are in a difference in that way, then that does also just, like, say something for kind of maximalism versus pluralism in the world. And I guess just one of the nice things about cryptocurrency as opposed to, say, traditional nation states is that it's just so much easier to kind of sort yourself into the community that reflects your values more. So I want to ask you about your opinion as to the scalability of these social organizations. that people find themselves in. So the way that I interpreted the convexity versus concavity graphs,

like one is a U shape and one is a N shape, right? And so if you put a ball on both of those, on the concavity, it rolls off to the edges, right? It falls off to the edge and then there's like a pit that the ball would fall into. And my mind goes to what Amin Soleimani calls a Moloch trap, right? A little trap where people like fall into where the convexity is, it's a valley, right? A rocks that fall down to the bottom of the valley, the center of the valley. So my gut take is that a convex worldview or convexity is a more scalable group of coordinated

people with the same opinions because they tend to converge upon pragmatism. They tend to converge upon the same ideologies, whereas the concavity seems to find itself in falling into a thousand different traps in different directions. Would you... Well, actually, and I think just one quick way to fix your analogy, unfortunately, I think it's the concave worldview, right? That's the one that's more inclined to stay in the middle. Like, the problem is that, like, in economics, the concept of utility is like positive utility is better, right? And so, like, the graph here is like how good. Whereas the gravity analogy is like, you know, you wants to fall down, right?

And so down as like the direction of the tendency or like preference if you wants to talk about fundamental. physical forces that way. So like look at it upside down, I guess, if you want to analogize it that way. But yes. And I think that, but like module of that, it's definitely a great way of looking at the issue. Because like I think one of the challenges with the convex worldview is basically that like if your worldview was that right now we're on the top of a mountain and we are going to fall on the left side of the mountain versus the right side of the mountain, then your first instinct is going to be, well, which side of the mountain seems less bad to fall down, and we're just going to fight like how to make everyone fall down that side as fast as possible

to avoid the greater evil of people falling down the other side of the mountain. And then if you have people on both sides, they're kind of both thinking that way, then, you know, you basically end up getting into a tug of war. So, and I think that, like, that's definitely a very kind of good consequences. of that intuition pump. So for those listening in who maybe haven't seen that yet, do we have some examples from the Bitcoin community or from the Ethereum community that kind of maybe prove this out?

I consider you almost in this article, like it's a little bit philosophy, but it's a little bit anthropology as well, right, where you're like studying the culture of Bitcoin and the decisions that the culture has made. and contrasting that to the decisions in the culture that Ethereum has. But what are some good examples of this? Like, was the block size debate a good example of this cultural temperament?

Absolutely. So basically, and if you remember in the block size debate, basically the block size at the beginning was one megabyte, and then there were all of these arguments about, do we make blocks unlimited, do we make them 20 megabytes? to make them eight megabytes, they would do some two four eight sort of thing. And eventually the side that ended up winning is very close to the extreme small block side. It basically says, I mean, blocks are one megabyte and then there's this one kind of soft fork called Segwit, which de facto increases the block size to 2.3 megabytes.

And then there's basically no increase that kind of going from there. Right. And if you remember during the block size debate in the big block camp, there were people who were in favor of moderate increases, so people who were in favor of, you know, eight but not 20 or four but not eight. And there were people who were in favor of just not having a block size limit at all because they felt that, you know, the system just needs to be able to support as much capacity as the means or like the market should be deciding block size. And like there's a bit of a tangent about how that particular claim is crazy because they

mechanism isn't really a market. But basically, like, you did have, like, these people who wanted very big. And I remember, like, one of the arguments that the very big block people would make is basically that, well, you know, if you oppose the, like, what the small block people are trying to do, then you should follow that thinking to the, quote, logical conclusion. And you should realize that ultimately the correct approach is no block size limit at all, right? And I remember being fascinated by that because like, okay, so basically if you're saying that, you know, moving from like point X to point X plus one, like if you support moving from X to X plus one, then you should also support moving from X plus 10 to X plus 11? Like, no, I don't believe that at all. Like, what kind of worldview is this? That's a convex worldview. That's a very convex worldview. it is indeed and so that was a convex people on the big block side right and then convex people on the small block side like they're like i argued against them in a lot of places like red it's whatever all the time right like a lot of the small blockers would say things like i i would say oh no you know see we need to compromise and we need to make sure that it's possible to read the to read the blockchain and it's also possible to write to the blockchain you need to send

the transaction transactions to the blockchain at affordable rates. And we need to find some way to compromise between those two. And they're like, no, either the system is decentralized or it isn't. And if it's 95% decentralized, then really that just means it's on a track toward not being decentralized at all. And either you can verify the thing or you can't. And everyone should be verified. Everyone should be totally verifying.

Yet anything that involves any degree of trust just is trust. And, you know, we should all just dismiss any level of trust as being basically the same as just using PayPal. And I remember also finding this worldview, like, very unreasonable and dogmatic as well. And actually, on that specific issue, I wrote another Vitalik.com article recently. If you remember the one that's called On Trust Models, that's was from August, right? Like, this is the one where I talk about the differences between, yeah, if you're a philosophy section, I think it's there. there, third one, there it is, so 2020, August 20. Basically, you know, there's N-of-N trust models, there's N over-2 event-trust models,

there's one of N-V-N trust, there's zero-event trust, scroll down a bit, and you see the graph, right, there it is. So, basically, the thing that I argue is that one-of-end trust is actually much, much closer to zero-of-end trust than it is to N-out-2 of-end trust, right? And, like, on the graph, like, it really does seem like it is much closer. a person who has a convex perspective on trust would say, you know, no, no, no, either you're at zero of N or your trust. And if you're trust, then you're on team PayPal. And, well, okay, PayPal is going to support Ethereum now, so we know, we can't keep PayPal.

I don't know who are we supposed to hate. What's interesting about it is because this convex bent, at first blush, it looks very strong because the community absolutely stands for things. There shall be only 21 million. Thou shalt not print more Bitcoin ever. But then when you like when you kind of look at it a little bit more and maybe this is just the concave in me kind of speaking, but you also see this brittleness.

Like I wonder, quite frankly, what the convex Bitcoin or community is going to do if it finds out that it needs more security than 21 million Bitcoin. at some point as these happenings continue into the future. Does it does does does that convex, uh, like mode just tear at the center and become brittle? Does the community completely fracture apart because it can't make any compromise on anything at all. It has to stick with sort of this, this maximalist type of ideology and philosophy. So it's weird because it appears that there's a strength to this like philosophy, but there's also a brittleness to it. And I'm not sure how to, I'm not sure

how to see that. And I definitely think that's a very good way of describing the tradeoff. Like this concept of shelling fences and just setting hard boundaries and saying we're just never, ever going beyond them. That's, like, that's definitely a powerful thing. It's a, it's definitely can be a scalable form of social coordination. It can be very emotionally appealing as well. But yes, It is brittle. And as soon as you enter into a kind of territory that you were not expecting, it can easily end up backfiring on you. So like in the case of Bitcoin, right, like if they discover the transaction fees are not enough,

they basically have three choices. Like one of them is print more than $21 million. The second is a switch to proof of stake. And the third is just accepts that the blockchain will get 51% attack from time to time. and you'll just have like people like sometimes days worth of transactions being reverted and it's surprised to be to an insane extreme extent just how many people are willing to accept the third option right right like I don't know have you seen any of this like even on Twitter I've seen a time of it like oh no no no no no you don't understand that like like transactions

being ever being final was never one of the intended properties of point. Pierre Richard makes this exact argument that it's no big deal, right? Like we wait days for transactions to finalize. Fitzhawk, would you characterize this as a Moloch trap? Good question. Is this a Moloch trap? I guess, well, a Moloch trap would be a situation where everyone or most people

want the situation to go in a certain direction, but just because they can't get along and ends up going in the opposite direction, I guess in that Bitcoin case, like, would it be a Moloch trap? Like one possibility would be that like people actually literally genuinely prefer the blockchain breaking once every few weeks to violating either one of the core tenants. But I think what would happen in reality is that there's a small minority that genuinely feel us that way. But the average bitcoiner, I think, is more pragmatic and they really would rather switch to proof of stake or at a small amount of issuance than. accepts that state of affairs. And so I could definitely see just like coordination failures between those two

leading to like a split or a lead or just taking a really long time to resolve the situation. And I definitely categorized that as being a mollog. Well, a split is not necessarily moleock. So Vitalik, it was very obvious to me when you were writing this that it was coming from the design philosophy of blockchains. the design philosophy of cryptoeconomic networks. But on the bankless program, Ryan and I talk about how these cryptoeconomic networks are very comparable to like a neo-nation, a new generation of a nation, right? And so when we're talking about design philosophy of a blockchain network, we're also kind of talking about the design philosophy of a new type of social organization, a new kind of nation.

And so when Ryan and I resonated with the concave view, it seemed to be out of an endeavor. and also when we kind of came to the conclusion that Ethereum seems to be resonating with the concave view, that Ethereum is trying to offer this credibly neutral layer. And what is credible neutrality other than being in the middle, being a pragmatist, about the middle of a curve? And so when a good nation state, to me, allows people to go about their day and achieve their personal goals without having the any sort of convexity of the design dispositions of a nation state or a cryptoeconomic network getting in the way of this individual's goals, right?

Like we want to empower the individual, but it's important to have an organization where the whims of one individual doesn't impact the whims of another individual in achieving their daily goals. And so that's where the neutrality of a nation state or the neutrality of a cryptoeconomic network comes into play in order to really empower the individual without enabling them to get in the way of empowering others. Was this kind of what you were thinking about when you wrote about this? Yeah, and I think this concave versus convex state definitely has implications going far beyond out of crypto politics.

And I think nation-state issues are definitely one place where this becomes also really a really important concern. Down the real line, I even give examples around nation-state issues like dealing with the coronavirus and taxation and how convexity and concavity and war, like how they end up kind of affecting those. And one philosophical thing that you could argue is that even the concept of freedom

is to some extent of fundamentally concave concepts, right? Because, like, one way to think about freedom is that like if side A is, my way of life is better and everyone should live according to my way of life and side B as your way of life is better and everyone should live according to your way of life then freedom basically says

well people who like my way of life live according to my way of life and people who like your way of life live according to your way of life and that is in some way a compromise between those two options right so you can definitely look at kind of concavity in that way but then

as we see there's a lot of kind of different layering here you can look at concavity in a lot of different ways. You can even look at concavity versus convexity as something affecting your personal decisions. So like a bit further above in the case of Bitcoiners, right? I mean the carnivory thing was one great example of this is like proof that this actually does get into like this is convexity, right? Like you're, you know, like a concave person would say, oh, you know, you should eat a balanced

diet and you should make sure to have like some, have some salad, have some good meat or fish or whatever you like, have some grains. But a convex person would say, I don't know, there's one thing that's healthiest and you should figure that out using research and you should just eat it, right? So carnivores are convex. Vecans are also convex. But on the other hand, normal people end up being more concave. And on the dietation, should I find myself more concave too? What a coincidence. It's funny because I think this, this is the type of article that you have to kind of like chew on and you can apply to so many areas of life. But I will tell you that one gift it gave me is actually understanding crypto Twitter and

understanding like Bitcoin or culture and Ethereum culture. Because for so long, Vitalik, like I just didn't understand it, to be honest. Like, I didn't understand the perspective where you'd kind of explain something or you talk about something and there'd be still such a chasm, such a delta. And it's because of these totally different world views. Like, once you see it, it kind of explains everything. I'm curious where you got this mental model of the world. I mean, have you seen anything like this elsewhere or were you influenced by anything when you were thinking about concave versus convex. I haven't seen anything quite that illustrates it in this way. Yeah, I'm not sure if like this exact presentation like existed in some literature before.

Like in the like concavity and convexity are definitely like things that you would understand in mathematics and like concavity is definitely a thing in economics. Like for example and like I talk about taxes below and that's like one of those examples there. but like I think it just like took years of brushing up against this kind of like A versus B logic and eventually just like figuring out like you know like yes this is where the divide is. So what are the implications here? Does this mean that at some level that Ethereum should try less hard to explain their positions to people who are very convex in thinking because there's just. going to be a delta there that can't be passed and maybe we should focus more on finding the other concave people in the world is that one implication what are your thoughts on the implications here it could be um i i think there's definitely people um who i think are just attracted to the the convex perspectives in a way that would be very hard to kind of make

get them to find the Ethereum camp appealing and you know maybe that's okay and again that's just like one property of like opposition to you know like Bitcoin maximalism but I guess there can be these

multiple ecosystems and they can all kind of prosper within their communities and I think ultimately you know for crypto to become mainstream like yes it does needs to just to kind of reach out to the wider world that you have to explain its perspective to the wider world more.

I do think that we should continue kind of making the case for Ethereum. And I think one of the ways in which I think the philosophy here can help is that I think Ethereum sometimes gets framed as not having values because it's not willing to just dogmatically take a direction and stick to it. But I think, like, I don't think that's not necessarily true, right? I think if you have values, it has things that it's pushing toward. But concavity is a good mental model to explain how it's possible to believe in something, but at the same time, be moderate about it.

And it's possible to actually have this as something you believe and not just like something you are because you just don't know what you are. So say we fast forward into 20, 30 years into the future and Ethereum is maximally successful to the degree that we all hope that one day it becomes. Do you attribute that success to a concave mindset on behalf of Ethereum? I definitely think the concave mindset is a very good one for at least what Ethereum's goals are, which I think involve a kind of broad adoption and bringing decentralization and permissionlessness to like basically large groups of regular people.

And on the Bitcoin side, it's interesting, right? Because I think Bitcoin is kind of of two minds in terms of what its goal is. Like one of its minds is basically replacing gold and like becoming the new digital gold. Whereas the other one is like let's be, let's focus specifically on censorship resistant it's wealth transfer for people in like authoritarian regimes or in general people and every regime is authoritarian to some non-zero extent. So like people who rush up against that side of things. And like having a convex approach, like it can definitely buy you a kind of success in that second

side of things because I know there really are people who just need fairly extreme tradeoffs. though at the same time I do think that Ethereum is also fairly capable of satisfying those tradeoffs except Ethereum's approach is different and it relies more I think on just

creating this thing that provides broad value for and of large groups of people even if like in some cases the average value per person could end up being a bit lower than you know literally saving your life savings and

like Ethereum can still literally rescue your life savings in some context, but the median use cases, I guess, a little bit more closer to average people's experiences. You know, there's an element, too. I wonder, I must in the Bitcoin culture community, if that second pool that you mentioned, that convex pool who has formerly been ardent supporters of, you know, censorship resistance, if that's not. being driven out a little bit too, because it just seems like this narrative of and this need for like price go up. Yeah, has driven out a lot of the original values. And maybe I'm

being a little harsh on Bitcoin or culture, but it just seems this embrace of kind of crypto banks, if you will, and like custody and institution. That wasn't part of the original convex plan for Bitcoin or so there's almost this hollowing out effect. where you get this convex culture around things like 21 million and number go up. Number go up and everything else. When our friend Nick Carter talked on Bloomberg about the having being priced in, and he was absolutely crucified because apparently that's not canon with some convex members of the Bitcoiner community, it's those types of things that have become more important

than elements of censorship resistance. And so there's almost like this hollowing out type of effect of the values that at least I thought used to matter in Bitcoin or culture. I don't know if you guys observed that, but it's something that has been surprising to me. These are not necessarily the extremists die for the cause martyrs that maybe they once were. I'm not sure. Maybe that's harsh. Yeah. And I like one of the kind of big fault lines I'm seeing is basically that so the store, the,

The story that small blocker values were originally sold under, right, is that you don't need that much security on the base chain because the lightning network can basically make up for it. And they brought a lot of their kind of heart and soul into the lightning network. But what's been happening over the last couple of years is that the lightning network has only received fairly limited adoption. And instead, we're seeing, you know, liquid start taking off. And liquid is, of course, a permission consortium chain. So basically the same sort of thing that they denounced as evil back in 24. 14. People are just, like, they're cheering on, you know, ETFs and, like, being the sort of institutional adoption that doesn't actually translate to any kind of decentralization. And, like, some of those

things, like, some level of those things is definitely kind of natural and good and, like, good for adoption. But it does seem, like, the problem for me, as when it gets to the point where it's so extreme that, like, these people are literally, like, start excusing the, like, the growing lack of accessibility of fully self-solverent options, right? Like, basically, like, when they start saying, oh, it's totally fine if, like, centralized intermediaries become used for most things, that's when it really starts, kind of turning me off. So it feels like it's kind of just letting the, like, it feels like it's the, like, it feels like it's letting the winds of kind of technical, a kind of historical default push it in a direction

and then retroactively kind of retconning the narrative to say that it always was that direction all along. Right. And like, there's definitely a big part of that that's kind of difficult for me to relate to. Like, you know, if they had, like, I think someone who is honest about maximizing censorship resistance would realize that intermediary minimization, is a pragmatic and needed part of censorship resistance.

And if you want to minimize intermediaries, you need to have the base chain be able to provide more things, potentially even at some level of risk to that making the base chain a bit heavier. But if you just care about permissionless system of the base chain, but you don't care as much about the ability to do trustless and permissionless things on top, then definitely gets the point where it starts to losing me. Well, Metallic, I really appreciated your post because it allows me to put so many of your other posts into perspective, right? Like, I understand your proof of stake posts better because I've read your concavity versus convexity post.

So thank you for producing this work. And also, thank you for coming on the bankless show to share some of that work in a new format to a new audience. And also, congratulations on the ETH2 deposit. contract Genesis as well. Thank you very much. It was good to be here. Thanks, Fatalik. See you later.