🎙 "It's Almost Like EIP-1559 Fixes an Economic Bug in Ethereum; More Usage Can Now Capture More Value:" Tim Beiko
In this week’s episode we speak with Tim Beiko, who organizes the Ethereum core developers calls and knows all the intricacies and latest development of the world’s most active blockchain. We spoke right after Beiko had proposed a date for the implementation of one of the biggest changes expected to happen to the Ethereum blockchain in recent history: EIP 1559. Tim explains in non-dev terms exactly how this change will reduce volatility in Ethereum gas prices and also reduce the uncertainty of whether user transactions are approved or not. Importantly, it also means that part of ETH paid per transaction will be burned, so that if demand for ETH and use of the Ethereum blockchain are greater than ETH issued, it should put pressure on the ether price to increase.
We also talk about the latest on the transition into ETH2, the timeline for the merge between the current Ethereum application layer and the proof-of-stake chain, and how Layer 2 solutions have made shards a nice to have and not a requirement for scaling. We also talk about Ethereum as a nation-state, and how radical economic experiments are its biggest exports.
The podcast was led by Camila Russo, and edited by Alp Gasimov. Transcript was edited by Owen Fernau and Dan Kahan.
🎙Listen to the interview in this week’s podcast episode here:
You’re a paid subscriber, which means you get the full transcript below. Subscribers also get exclusive access to The Defiant’s Discord chat for the community, here’s a new link to join.
Camila Russo: So I usually start the show by asking guests to tell me about themselves and their background. But because we have breaking news that Tim proposed a launch date for EIP 1559, I want to start with that. Can you explain what EIP 1559 is, why is it important?
Tim Beiko: So yeah. Quick overview of 1559. It's a pretty big change to Ethereum and that's one of the reasons why people talk about it so much, because it affects a lot of things. But at a very high level, what it does is it makes it explicit in the protocol what the amount is that you should pay for a transaction. So right now, when you use Ethereum, everybody sets their gas price, but the protocol doesn't tell you what's the right gas price, right? The way that this happens is your wallet, say you're using MetaMask or My Crypto or something else, they'll look at all the previous blocks based on the data and how much people paid in those blocks. They'll try to come up with an estimate and they'll put that gas price on your transaction.
“Quick overview of 1559. It's a pretty big change to Ethereum and that's one of the reasons why people talk about it so much, because it affects a lot of things. But at a very high level, what it does is it makes it explicit in the protocol what the amount is that you should pay for a transaction.”
But very often that will be wrong in a way, right? Like, we've all had this experience, you set a transaction on MetaMask, it tells you it's going to confirm in two minutes, then you refresh the page, and now it's seven minutes. So then you're speeding it up. It's just a very clunky process. And one of the reasons for that is, like I said, there's no notion of the correct gas price in the Ethereum protocol today. So what 1559 does at its core is it adds this to every block. So every block will have a new value called the base fee per gas. And this base fee is the minimum that you need to pay for your transaction to be included in that block. So it makes it very easy that you can look at a block and say, okay, this the minimum price I need to pay.
So I'll just put that as my transaction fee and get my transaction included. There's a lot more nuance and subtlety, but that's kind of the gist of it. The first question people ask about that, it's like, well, you know, that sounds pretty simple. Why can't we just have that already? One big challenge in gauging what's the right gas price to give users is we need to figure out what's the demand for Ethereum block space right now? Because the supply of block space is pretty fixed. We know we have one new block every 15 seconds. The gas size is fixed every block. And right now all these blocks are full. So if you just ask the network what's the demand, it'll tell you a hundred percent, and that's not a very useful indicator.
So what we do with EIP 1559, the other big change is we make blocks twice as big. So say right now we have 50 million gas on main net. We'll expand it to 30 million and we'll aim to keep them only half full. And when they're more than half full, we'll just raise this minimum price. Kind of like Uber's surge pricing in a way, where if everybody wants to get an Uber at the same time, the price goes up to discourage people. But then the opposite is also true. So if we're less than half full, we'll just start lowering the minimum price. So this is the way by which we can gauge demand. And it has a nice side effect in that most blocks will always have extra room. Right now blocks are basically always full.
“Kind of like Uber's surge pricing in a way, where if everybody wants to get an Uber at the same time, the price goes up to discourage people. But then the opposite is also true. So if we're less than half full, we'll just start lowering the minimum price. So this is the way by which we can gauge demand. And it has a nice side effect in that most blocks will always have extra room.”
So when you set a transaction, you're kind of waiting. And if somebody bumps in line in front of you with a higher gas price, then you're waiting behind them. But with 1559, if you assume the average block is 50% full, it means when you send your transaction, the next block should have some spare room to include it, so you should be included right away. And so it should help the average users, assuming you're willing to pay the current price on the network, to get their transaction included much quicker. The last part of 1559 is the fee burn. That's what a lot of people focus on. But the reason for the fee burn is because we have this minimum fee that can go up and down, and we want to use this as a reliable indicator for demand on the network.
If we did not burn the base fee and we just sent it to the miner of the block, what the miner would do is they would fill the block with their own transactions so that it would raise the fee for everybody else, but then they wouldn't actually pay this fee because they would get the transaction fee back. So the original reason to have the fee burn as part of EIP 1559 is so the miners are indifferent. Whether the minimum price is high or low, they're going to get a small part of the transaction fees for every transaction. But it shouldn't be higher or lower based on the base fee on the network.
Fees for Miners
CR: So do miners get a fixed fee.
TB: Basically right now, transaction fees on main net represent three things. They represent, one, you just getting included in a block, two, how early you are included in the block, and three, your competition against everybody else who wants that. So after 1559, miners will not get this base fee. So this base fee is burnt. But every transaction would include a small part of the fee that goes to the miners. And that basically compensates miners for running nodes, for the risk of having basically [inaudible] blocks. So if miners got zero fees at all, they would just mine empty blocks and not even bother maintaining transaction pools and propagating bigger blocks on the network. And the other part of fees that miners will still get is everything related to wanting to be the first in the block.
So these are the fees related to MEV or miner extractable value. When people have, say, an arbitrage transaction that they want to do on Uniswap, they're willing to pay a lot of money to get that transaction included first in the block because otherwise that arbitrage opportunity is gone. And so they'll still be willing to do that even after 1559, and miners will still be able to collect all of those fees. So it's really just the minimum fee per transaction to be included in a block that gets burnt. But then the other part of the fees goes to miners.
CR: Right. And so does that mean that 1559 doesn't really have an impact on MEV, on minor extractable value?
TB: No, it doesn't. Teams that work on MEV will have to adapt to 1559 and add support for it, but it won't reduce or increase MEV.
CR: Okay. And what percentage of total fees is the base fee that miners won't get anymore?
TB: So there's no exact value yet. There's been different estimates and it's probably the one thing people are eager to see after after main net. The estimates range from 25 to 75%. So it's like a really wide range that I've seen. And it really depends. The challenge with those estimates is it depends on what period you look at. So if there's a period where not much is happening on-chain, then there's probably not going to be a ton of fees going to miners. But if there's a period where there's a bunch of things happening, like a bunch of liquidations or some new DeFi protocol launches, everybody wants to get in, then those periods where there's really high activity, the fees for miners will be quite high at that point because people will be willing to pay high tips to miners to get included. That's why it's so hard to tell, because you need the average to say, well, how often do these periods of very activity happen? When they happen, how much are people paying in total and, based on how frequently they happen, what's the average over time? So we don't have those numbers yet.
“So if there's a period where not much is happening on chain, then there's probably not going to be a ton of fees going to miners. But if there's a period where there's a bunch of things happening, like a bunch of liquidations or some new DeFi protocol launches, everybody wants to get in, then those periods where there's really high activity, the fees for miners will be quite high at that point because people will be willing to pay high tips to miners to get included.”
CR: It seems like this is a pretty negative change for miners. Like how was it that this is going through?
TB: Yeah. So, I'd push back a little bit on pretty negative. Obviously, it does reduce some of miners' income, and that's pretty clear. But it's worth noting 1559 is not the only thing that affects miner revenue. Transaction fees are one part of it. But the two other big parts are the block reward and the difficulty. And so the block reward and the transaction fees are denominated in ETH. So obviously the value of ETH against the dollar matters a lot because miners' costs tend to be in the real world. So they tend to be denominated dollars. So if they get a two ETH block reward at 4,000 ETH, it's twice as good with regards to their costs as at 2,000 ETH.
And then the other thing is the difficulty on the network. How many other miners are mining at the same time? So with a given set of hardware, if there's more people mining, you're going to make less money. And if people go away, then you're going to make more money. And so these two variables affect mining revenue to a much greater extent than transaction fees. And especially if you look at the entire development history of 1559, when it was first proposed transaction fees on Ethereum were basically zero. So this wasn't a major concern and the original version of the EIP actually had that listed, saying this won't be a big deal because miners make almost nothing in transaction fees anyways. But then last year, when we were in the middle of the development of it, there was probably the most minor revolt when DeFi summer was happening and fees were just crazy high.
And I think then, miners were looking at that, then saying, well, if this cuts 75% of the fees, the fees are almost bigger than the block rewards, then that cuts half of our income. But the fees have dropped since then, right? The fees have dropped since they were at 500 gwei. I think today they're at a hundred and people think that's high. Just last week, they were like sub-10 gwei. And 1559 is not live, so clearly miners are able to adapt to that variability in fees. The last thing is, if you look at the needs of the Ethereum network, we want the network to be secure.
“So I don't think it'll be that big of an impact at the end of the day. But it definitely felt much worse when gas prices were 50 times higher than they are today.”
That's the number one thing. And the security needs of the network don't change based on the gas price. If we think that the network is not secure enough, we should probably do something like raising the block reward because we'll get more miners to come in and there'll be a more predictable source of revenue for them, then relying on transaction fees. Because as we've seen, transaction fees can go up and down. And if we're only secure on the network at 200 gwei or more transaction fees, and we go below that, then that's a terrible place to be in. And I think as we're seeing now, the network could operate fine at sub-10 gwei prices. So I don't think it'll be that big of an impact at the end of the day. But it definitely felt much worse when gas prices were 50 times higher than they are today.
Blocks Half Full
CR: Great points. Another question on the explanation that you gave earlier on increasing the gas price per block to make it so that each block is half full. Could it just happen that with time, as demand for Ethereum block space increases, that blocks get full again?
TB: That's a great question. Basically the way it works is the base fee goes up very quickly. So it goes up about 10 X in five minutes of full blocks, and that's exponential. So in 10 minutes it'll go up a hundred X, in 15 minutes it'll go up a thousand X. So that means that if there's a ton of new demand, say, Uniswap launches a governance token and everybody wants to go claim UNI, when that happens, in that short period prices will go up a ton because everybody wants to use. But at some point, once we're at 500 gwei or 750 gwei or a thousand gwei, there's not that many people that want to do transactions at that price. So they'll stop making transactions and then slowly we'll get 50% full blocks again.
“So what happens if there's more demand, the base fee will just increase up to the point where there's about 50 million gas of demand at that price. And people who want to pay more, they just will pay that amount and get a refund for the rest, and people who don't want to pay that amount will just have to wait for it to come back down.”
And because the speed adjusts so fast, in 10, 15 minutes, we can get a thousand-ish X range in gas prices. It's unlikely you see cases where there's more than that duration of full blocks because then you'd be at 10,000 X, or 100,000 X increase in gas price. To give your listeners an idea, in the whole history of Ethereum, we've seen about like 1,000 X range between the lowest and highest gas prices. So what happens if there's more demand, the base fee will just increase up to the point where there's about 50 million gas of demand at that price. And people who want to pay more, they just will pay that amount and get a refund for the rest, and people who don't want to pay that amount will just have to wait for it to come back down.
CR: Got it. Okay. So on today's announcement, what does this mean? So if you can walk us through what's the status of EIP 1559, how it's been on Testnet, and what does this launch on August 4th mean?
TB: Yeah, so 1559, as we've been discussing, is a pretty big change to Ethereum. So we've tested it very extensively for the past year. We've run analyses on it, run up some new testnets, spammed those testnets, found bugs, fixed them, and so on. So it was really a long process to get it to a spot where we could even consider deploying it for maintenance. Late this winter, early this spring, we decided to add it to the main net. And then even there, as we were finalizing things and whatnot, we wanted to be sure that deployment on Testnet was smooth before we even thought about scheduling a date for Mainnet. So we scheduled the first three testnets.
And now we've had the first two go successfully. So we first had Ropsten and when Ropsten happened we basically spammed the network for a day to makes sure that the base fee would go up and come down and that we could handle large blocks on the network. Then we had the Goerli fork last week where we spammed the network before and after the transition. So just to make sure that everything works smoothly. So now we're in a spot where we're pretty confident it works, nothing broke. And so now core developers were comfortable agreeing on a date to deploy us on Mainnet.
So this morning I proposed August 4th. It's still not a hundred percent final because developers were across different time zones. I wanted to leave at least 24 hours for any team to say if they're not ready, like if a team doesn't feel like they'll be ready by then we should obviously change it. So far, nobody has stepped up and said that. Assuming that doesn't change tomorrow, then yes, we would be looking at August 4th, for it to go on Mainnet.
Prepping for the Future
CR: So a huge change on Ethereum is just a few weeks away. What should Ethereum users, ETH holders, ETH developers, expect from this? How will they be impacted?
TB: Sure. I guess the first thing that's worth noting is 1559 is opt-in in a way, so legacy will keep working on the network. And they'll, kind of behind the scenes, be converted to 1559 transactions. So the base fee is still burnt in them. So they don't get around the mechanism, but if you're using your wallet or you are a wallet and you're looking at this like there's no way we can be ready for August 4th, that's totally fine. Legacy transactions will keep working. And that's something that was really important as we were developing this, because it can take time for people to upgrade. Or, imagine you've already signed a transaction and, you know, you want to just broadcast it after the fork, that will also work. So for most people, you basically have to do nothing unless you want to.
“So they don't get around the mechanism, but if you're using your wallet or you are a wallet and you're looking at this like there's no way we can be ready for August 4th, that's totally fine. Legacy transactions will keep working. And that's something that was really important as we were developing this, because it can take time for people to upgrade.”
If you're running a node on the network you're going to have to upgrade to a node that has the London network upgrade. We don't have the client releases out yet. They should be out next week. There's going to be a blog post on the blog, the ethereum.org website. So you just need to download the new version there and that will make sure that you stay in sync with everybody after the upgrade. Similarly, miners will have to do that, exchanges and whatnot. Yeah. That's pretty much it. I mean, if you are a user or smart contract developer you can use 1559 through any of the wallets that will support it. I know My Crypto and MetaMask are working quite hard on this right now. I'm sure others are too. It's just those two I've I've been talking with.
They're not the only ones, but the ones I'm aware of. And also for smart contract developers, one thing that's really nice is alongside 1559, London will also bring a new upcode in the EVM that returns the base fee for a certain block. So if you're doing things like automatically generating transactions in your smart contracts, stuff like that, you'll be able to look at the current block or the previous block's base fee and set the price for your transaction automatically based on that. One last thing that's also worth noting, if you are a developer or even a user, and you want to try this, it is live on Goerli and Robston right now, so you can go on those networks and play around with it and experiment.
CR: For end users, what will it mean in terms of gas price? Should people expect lower gas prices, or you mentioned that it might get pretty volatile at times?
TB: It won't drastically lower gas prices. Because gas prices are a function of supply and demand. And so, you know, the supply of block space on Ethereum is pretty fixed. And the demand varies a lot. So if a lot of people want to use Ethereum, the gas prices will go up. There's nothing you can do about that except stuff like layer 2. Basically, it's creating more supply of block space.
CR: So what should people expect?
TB: With regards to the gas prices, it won't lower the gas price, but it'll make it clearer what the right price is. And one way to make that a bit less abstract is if you look at blocks right now on Etherscan, and you look at say, what's the lowest 10% price that somebody paid and what's the median price? Sometimes there's like a five to 10 times difference between that. So that means that two people in the same block, literally somebody paid five times more than the other. But their transaction gets executed at the same block, and this excludes MEV. So what it'll mean is that everybody who gets into the same block pays roughly the same price.
One kind of non-blockchain analogy for that that, one of the researchers working on this did, is right now, if you're making a transaction on Ethereum it's kind of like buying a house in a hot real estate market where the house is listed at a million dollars.
“After 1559, it's more like when you buy something in a store where it'll tell you, look, the minimum price for those blocks is X. And if you want to buy it, if you want to have your transaction included at that price, you can just pay it. And otherwise, you just don't. But at least you know that if you paid, you've paid the fair price.”
You know a lot of people want it. So like, you're going to put $1.1 million. I put $1.2 million, and somebody else puts $1.5 million. And they get it, but they overpaid by like $300,000 because they could've just basically put the second-highest price and still won. Now transactions on Ethereum are basically like that, but it's not only the person who buys it that pays, it's like everybody included in the block. Whereas after 1559, it's more like when you buy something in a store where it'll tell you, look, the minimum price for those blocks is X. And if you want to buy it, if you want to have your transaction included at that price, you can just pay it. And otherwise, you just don't. But at least you know that if you paid, you've paid the fair price. And if you choose to not pay, you know what the price was. So that's the effect that it will have on gas prices. It's a bit more subtle than whether it'll raise or lower them, it just helps reduce the variance within a single block of the gas prices
Getting into ETH
CR: Perfect. I love that analogy. It makes it super clear. Okay. Now I think it's time to get more of your background after you explained in great detail everything about 1559. I'd love to just learn more about yourself. Where did you come from and what got you interested in crypto in the first place?
TB: Sure. Right now, I run the core developer calls. So these are the calls where developers of the Ethereum protocol get together to decide what changes will happen to the protocol. I've been working in this space for three and a half years right now. Before doing this, I was a product manager on one of the client teams. So as part of Consensys, I worked on their base client. And before that I was also a product manager, but just following the space a bit more as a hobby. I first heard about Bitcoin in, I think it was 2013 or 2014 from a friend. Seemed weird at the time. So didn't do much about it, unfortunately. And then a couple of years later when the DAO was about to launch the project Ethereum, he mentioned Bitcoin again and Ethereum.
So then I got interested. I bought Ether literally the day before the DAO got hacked to invest in the DAO. And that was kind of a wild onboarding experience because right after that there was Ethereum Classic, and you had to split your tokens to make sure that you wouldn't get your transaction replayed on the other chain. I knew nothing about Ethereum other than like the name. So it was quite a stressful experience. And after that, it kind of felt like things died down pretty bad for like a year. So I kept following, but frankly, not that much was happening. And I started paying attention again, I think it was like late 2016, when you started seeing the first projects have token sales on Ethereum again. It felt like at least there was some renewed interest in the platform and it hadn't completely died.
“But I was like, you know, even if all of these were scams in a few years, there will be some other use case for this thing. And when there was the ICO's in, in 2017, it just grind the network to a halt. Like it just showed how brittle the protocol was. So that's when I decided I wanted to work at the protocol layer.”
And then in 2017 there was obviously this huge ICO boom. And what I found interesting was, it showed there was tons of demand for Ethereum. Like even though all of the ICO's could have been scammed, and it turns out, I think people were probably too harsh in criticizing them in 2018, 2019, a lot of them today are actually working. But I was like, you know, even if all of these were scams in a few years, there will be some other use case for this thing. And when there was the ICO's in, in 2017, it just grind the network to a halt. Like it just showed how brittle the protocol was. So that's when I decided I wanted to work at the protocol layer. It just felt like there was a lot of demand, a lot of work to do. It took I think a year to go from just contributing part time and learning to get a full-time job in this space. And I've been doing that ever since.
CR: Very nice. What do you think attracted you to keep following Ethereum, even after the DAO, even after it seemed like there was nothing going on.
TB: So there's a few things. Like one, it's funny, blockchains are like a category today. But I think when Ethereum launched, you know, it was Bitcoin, there was a bunch of other random coins, but Ethereum was really kind of this unique thing, like, it was kind of hard to categorize. It was quite different from Bitcoin and Litecoin and Monero. So I don't know, just some interest in the fact that it was such a novel experiment. I think the community, the quality of the Reddit, their research updates, even that was quite good. So yeah, it just seemed very interesting. And, you know, I think this is also a pretty common experience. When you buy the token of something, you kind of feel more invested in it. So, Ether crashed, I think it was like 75, 80% after the DAO. And at that point, it didn't even feel worth it to sell anymore, but it was like, well, I guess I'll just keep following it, like who knows? So I think a combination of those three things, just like the technical interest, the community, and the fact that you're already invested in it.
1559’s Burning Mechanism
CR: Very cool. And here you're kind of at the very core of it. So I also wanted to ask you about another key part of 1559, which we briefly went over. But this burning mechanism, what does this do to ETH, like the asset?
TB: Yeah. Basically, at a very technical level, what it does is every time there's a transaction, it burns a little bit of ETH, and the more people want to transact, the higher the base will be, the more ETH gets burnt, right? So from a very objective level, that's what it does. And you know, people like to talk about the impact on the price and whatnot. And it's always a bit hard to talk about that because there's so many variables that affect something like the price. So if you work to isolate every single variable and compare, Ether with the burn and Ether without the burn, obviously with the burn you're going to get a lower supply. And if there's the exact same demand for Ether, the price should be higher.
“It creates a positive pressure on the price because it reduces the supply if you assume that the demand is constant. But that's a very, very risky assumption because so many things influence demand for that.”
But you know, if I say that, and the price today is like 2300, something like that. I don't want people to buy Ether thinking well the burn will make the price go up, because tomorrow, you know, the U.S. could ban Ethereum and the price could go to $500 even if we burn a bunch of Ether. So I think, you know, it is like a pot. It creates a positive pressure on the price because it reduces the supply if you assume that the demand is constant. But that's a very, very risky assumption because so many things influence demand for that.
I guess the way I like to see it, it's almost like it fixes some economic bug in Ethereum, where if I can tell somebody who knows nothing about blockchain about Ethereum, they'll assume that the more usage there is on the platform, the more the platform will capture value. They maybe won't be able to say how, but it's kind of an assumption you'd make, and today that's not true. It's like more people using Ether does not really have any impact directly on the network value. Whereas, if you do burn more and more as more and more people use the network, then it does create this positive feedback loop. And I think it's really valuable. It feels like the longer term you look at it, the more you can probably draw a correlation because if you assume that the usage keeps growing over a long period of time, then the total amount of burn will be much higher, and it all represents a meaningful part of this supply.
Two more things to consider is, is one, the price of Ether is interesting just by virtue of being the price. But it's also like economic bandwidth for Ethereum because when Ether is used as collateral, it's like the higher the network value, the more loans on MakerDAO can be opened that are denominated in dollars. And the larger collateral can grow and the more you can support a big economy on Ethereum. So I think that's the part that I want to optimize for. It's like how, over the long-term, do you align with Ethereum's growth with a thriving economy on the network?
“...the price of Ether is interesting just by virtue of being the price. But it's also like economic bandwidth for Ethereum because when Ether is used as a collateral, it's like the higher the network value, the more loans on MakerDAO can be opened that are denominated in dollars.”
And the last part about the burn, people talk about deflation and the impact on that. So it's worth noting that, you know, in order for the burn to lead to deflation, you need to have the base fee be higher than the block reward for whatever period you're looking at. And specifically that excludes the miner tip, right? Because that part of the transaction fee doesn't get burned, it's not going to contribute to deflation. So when people are trying to estimate that, I think it's important to look at, well, obviously what's the inflation today and how is it going to change? And with the merge from proof-of-work to proof-of-stake, it'll reduce by, I think it was something like 8X roughly. So it'll obviously get much lower. But we also need to be mindful that if there's a medium-high base fee and, you know, high tips, that it's only that base fee that gets burnt and contributes to the deflation and not the entire transaction fee.
“...we also need to be mindful that if there's a medium high base fee and, you know, high tips, that it's only that base fee that gets burnt and contributes to the deflation and not the entire transaction fee.”
CR: Okay. So there's many different variables to consider before saying burning Ether because of EIP-1559 directly will contribute to ETH price increasing or to ETH being deflationary. Many other things will also have to happen, like sustained increased demand.
TB: And increased demand is important for two things, right? It's important obviously for people buying the asset, if you care about the price, but it's also important if you want to have a high burn. Because you know, basically what gets burned is the function of the demand for the network. So I think the burn is a nice marginal improvement, but if there's no demand for Ethereum, it won't do anything.
CR: Makes sense. So to put a bit of numbers to this right now, and correct me if I'm wrong, the last time I looked this up, Ethereum inflation was around 4% annually. Is that right?
TB: I know the block reward, but I don't know the yearly inflation based on the supply.
CR: Okay, okay. Cause I was going to ask if that's the current inflation rate, if there are any estimates…
TB: What I did calculate is how high does the base fee need to be in one block for that block to be deflationary. And so on proof-of-work, it's roughly 150 gwei. It's a bit less if you don't count uncle rewards, but if you assume some amount of uncle rewards, if the base fee is over 150 gwei, and we have roughly 15 million gas, it should be a bit over two ETH that gets burnt at that block. So that's a rough proxy. But when we move to proof-of-stake, it's much lower. It was somewhere between 15 and 30 gwei. So because of proof-of-stake you're issuing much less to validators than you are to miners on proof-of-work. The level at which base fee has to be in order for the system to be deflationary is much lower.
“...if the base fee is over 150 gwei, and we have roughly 15 million gas, it should be a bit over two ETH that gets burnt at that block. So that's a rough proxy. But when we move to proof-of-stake, it's much lower. It was somewhere between 15 and 30 gwei.”
CR: So with gas at a hundred gwei today, that would mean on proof-of-stake ETH would be deflationary at least at today's prices?
TB: Yeah, exactly. And you know, on proof-of-work, you wouldn't be fully deflationary, but say, assume my 150 number is right and gas is at a 100 today. Then it means you basically burn two thirds of the block reward. So 1.3-ish ETH. So you would still reduce it, right? So there is always a deflationary pressure, but to go from being, you know, net inflationary to net deflationary, it's roughly 150 and 15 to 30 after the merge.
Changing Ethereum’s Architecture
CR: Super interesting. Okay. And then now that you mentioned the merge this is another huge milestone that ETH will undergo in the next few months. So if you can take a step back and explain what the merge is about…
TB: Sure. So it's worth taking one or two steps back because a lot of people are still very familiar with the old phase zero, phase one, phase two, Ethereum 2.0 roadmap. That's a good point to start to explain how that's changed. So phase zero has always been the beacon chain. So just having a proof-of-stake engine, that's not attached to the current mainnet chain. So we could just test it in production because that is the most complex part of the system. And then the original roadmap would just add shards to this proof-of-stake engine, but not have the shards run computations, just have them store data. And then phase two would transform these shards from shards that store just data to shards that can actually run the EVM like the current Ethereum mainnet.
So the teams were working on this. They got phase zero shipped last year in December, if I recall correctly and then started thinking about, okay, what should we do next? And at that time it started to become clear that rollups might be actually a better way to scale the computation on Ethereum than sharded computation because, you know, we could get them live on mainnet before the whole proof-of-stake transition was done. It also doesn't require the protocol to decide what the rollup does. Right. Like any company could go and create their own rollup implementation. They don't need to come to all core devs to decide if it's okay. And it's not changing the protocol. So it gives power to the community.
“...at that time it started to become clear that rollups might be actually a better way to scale the computation on Ethereum than sharded computation because, you know, we could get them live on mainnet before the whole proof-of-stake transition was done.”
So as we started seeing that, there was this shift in the roadmap from saying, well, instead of having like all these shards that we first stored data on and then turned to computation, maybe we could just use rollups and those can be the shards that run computation. And if we did that, then we could move to a system where maybe the only computation shard is the current Ethereum mainnet. We don't have to change anything about it. We can add rollups to it for scaling, and we can just link it with the beacon chain in order to remove proof-of-work. So that was like the first iteration of the roadmap. And then after that, there were two other changes. The first was to say, well, right now, Ethereum clients like Geth and OpenEthereum, and Nethermind, they're already used to using different consensus algorithms.
So they use proof-of-work on mainnet, but on testnets like Goerli they use proof-of-authority. So one researcher on the consensus team started thinking, well instead of having, mainnet as a shard, that's linked to the beacon chain, why don't we just combine them together and say, well on this block use proof-of-work and then on this block, just start using proof-of-stake and have that be your new consensus mechanism. And that would just make it much simpler in terms of architecture. And so they started prototyping that, and last, I think it was in May, there was a big hackathon where they actually managed to run these kinds of combined clients of like a Geth or a Besu, or a Nethermind that's running the EVM with a Prysm or a Lighthouse that's running the beacon chain and that's serving up the consensus mechanism.
And so this is kind of the design that we're going for for the initial merge and allows to keep basically all of the ETH 1.0 clients. There's still significant changes, but we're not rewriting something from scratch. We're just taking what's there and adapting it, so that it can communicate with the beacon chain easier. And that's what we're going to try and do first. This way the operations of mainnet won't be disrupted, rollups will still be a way that people can use the scale. And once that's done, we can add shards just to store data because rollups end up generating a ton of data. If we have data shards, it'll be just much cheaper for them to store the data and will be cheaper for users to use. But it goes from sharding being kind of a must-have to a nice-to-have just to reduce costs. So that's the goal. Like I said, we've had prototypes working in May and now both teams on ETH 1.0 and ETH 2.0 are a bit busy with London and 1559 on ETH 1.0 and Altair, which will be the first hard fork on ETH 2.0. But basically as those wrap up the merge will be the focus of both ETH 1.0 and ETH 2.0 teams.
CR: Very cool. Okay. So to summarize what's expected to happen is that in the next few months, I mean, I've heard different estimates, end of this year, Q1 of next year, what will happen is that Ethereum clients, for those listening, are what companies that run Ethereum software, It's like the way to access Ethereum?
TB: Yeah. They're like the nodes on the network. So instead of running a single node, it'll still be like a single software package. One way to think about it is, say you're running Geth or Prysm today, you'll be running a combined version of like Geth or Prysm or Nethermind and Lighthouse and that'll be your Ethereum node. And at the transition, it'll be a bit weird because it needs to still have proof-of-work and have proof-of-stake. But once the merge has happened, you'll just have your part of your Geth node without proof-of-work, we call that the execution part, and then your say Prysm or Lighthouse node which we call the consensus layer. And they'll just be running together as a single piece of software.
“...once the merge has happened, you'll just have your part of your Geth node without proof-of-work, we call that the execution part, and then your say Prysm or Lighthouse node which we call the consensus layer. And they'll just be running together as a single piece of software.”
So everybody who's running a node on the network will need to change. And, you know, there'll be packages where you just download one thing and they're all combined together. But one thing we do want to try and maintain is the ability to mix and match so that if there's like four ETH 1.0 Clients and four ETH 2.0 Clients that you can choose the one of each that works the best for you. And the merge will just be saying after this block on proof-of-work, then get the next block from the beacon chain and start adding transactions to that block.
CR:Okay. So, so basically what will happen is that you will have clients or Ethereum nodes running the Ethereum software which will be dedicated to the execution layer of Ethereum, the computation layer, like everything that is made to be able to run all the different Ethereum applications and run smart contracts and all of that. And then you will also have other Ethereum nodes running Ethereum software that are dedicated to running the proof-of-stake consensus algorithm instead of the proof-of-work consensus algorithm that's being currently used. And so when you combine those two, like nodes running the operational layer and nodes running the new proof-of-stake consensus layer, you will get the new Ethereum running on proof-of-stake without having to change any of the current applications that are now working on Ethereum.
TB: Yeah, exactly. And there's going to be some small changes for applications. Just there's things that like, for example, there's an opcode that returns the mining difficulty right now in Ethereum proof-of-stake has no difficulties. So, you know, there'll be some small changes where we need to fix those features, but as a general rule, Ethereum won't won't change.
“...there'll be some small changes where we need to fix those features, but as a general rule, Ethereum won't won't change.”
No Dates Set for The Merge
CR: Okay. And then are there any updates on dates for this merge?
TB: So I think we're still pretty early for dates. You know, I think it'll be our main focus, right around August as we wrap up with London. I'm very cautious of saying dates because the risk is we still don't know what we don't know. So with the information I know today I can say oh yeah we'll get there in Q4, but maybe there's a thing that's broken that we don't even know about yet. And that'll delay it for two months. The one thing I am pretty confident in is, I was a bit skeptical that this bundled client architecture would work well. And the fact that it may, we actually manage to get it up and running as a prototype in less than a month, made me much more confident in it because, there might be a bunch of engineering issues we need to figure out, but the general architecture works. But yeah, unfortunately I feel like it's too soon to be confident in the date.
“I'm very cautious of saying dates because the risk is we still don't know what we don't know. So with the information I know today I can say oh yeah we'll get there in Q4, but maybe there's a thing that's broken that we don't even know about yet. And that'll delay it for two months.”
Withdrawing Staked ETH
CR: Makes sense. Okay. And then when the merge happens, will the main change for end users, be the ability to withdraw staked ETH, because right now the beacon chain, the proof-of-stake chain is live. People are staking ETH on the proof-of-stake chain, but you can't really withdraw ETH or do anything with that ETH. So is that the first thing people expect?
TB: So that won't be at the exact merge point. And the reason why is that the merge is already complicated enough that we just want to switch from proof-of-work, to proof-of-stake and make sure that works and not bundle any other changes at the same time. That'll be the first feature after the merge. So people are calling it like the cleanup for it. So obviously withdrawals are the most important things. And there's a bunch of other small things that we want to do shortly after the merge. And it just feels like doing it all together, risks delaying everything longer. Whereas if we can just get the merge tested and shipped then enabling withdrawals becomes a somewhat smaller change. Whereas if we do it all together, you're just adding all this complexity at once and that's just harder to test.
“...if we can just get the merge tested and shipped then enabling withdrawals becomes a somewhat smaller change.”
CR: Okay. That makes sense. And so how much later would withdrawals be enabled?
TB: Like a few months, I suspect, definitely not a year. You know, usually our hard forks happen between three and nine months. I think this one people will want to see it happen pretty quickly. So I suspect it's six months or less after the merge. Obviously that assumes the merge goes smoothly. Right. If there's a big problem with the merge that'll delay. But yeah, it'll be priority number one. And I think also, people are aware that there's a desire to withdraw to staked ETH and so it'll be the first thing we want to enable.
Rollups and State Expiry
CR: And it's so interesting what you were saying before about roll-ups because it really completely changed the phases that Ethereum developers had always been thinking about. It used to be, okay, first proof-of-stake, scaling comes next, but I guess because of the needs of Ethereum users that people needed scaling now. So I guess that pushed for development of all these different solutions.
So it's really interesting that instead of having to rely on, on shards, which would come, I'm guessing maybe years after the merge, at least with like computation on them and all that, we have right now, rollups, which are providing those scaling solutions or at least on their way to with Optimism and Arbitrum, as the two main ones. So do these completely replace sharding? And I'm wondering why the focus is on just rollups because there are so many different other layer two solutions. So is there any reason why, from your side, that's the focus?
TB: So I'll start with that question because it's easier. I think rollups are a nice combination in that they provide the security guarantees that are similar to mainnet. And they work. They're live today. So it's high security. And I wouldn't say low engineering complexity, cause they're pretty complex, but relatively low compared to other options. And even within rollups, you know, there's all of these debates between optimistic rollups versus ZK rollups and there are different approaches with different degrees of technical complexity, but at a high level they work and they're secure. So it makes a lot of sense to use them. I think with regards to computation shards it's definitely something we can do in the future. It's for sure, a couple years out, there's a bunch of unsolved problems with it.
And I think more importantly there's just other more important things on the Ethereum roadmap. I'd say the biggest one after proof-of-stake is this idea of state expiry. So the biggest challenge on Ethereum right now is that all the applications are stored on the network forever. And so that means that when you, when you create a smart contract or when you add data to a smart contract, say you mint an NFT, it creates a new one and it maps it to your address. All that data is stored on the network forever. And so it means every node that joins the network has a bigger and bigger amount of active data that they need to store. And that just doesn't scale to, you know, infinity. What newer blockchains do for this is they have this concept of rent where instead of your CryptoKitty or your other NFT being on-chain forever, it's like you put a bit of ether and if you run out, you know, it just gets deleted from everybody's node.
And you can keep a copy yourself and you can revive it, but it doesn't impose a burden on everyone. Because Ethereum is already live and we already have contracts on everything. If we added rent tomorrow we would break every single contract. That would not be great. So we need to come up with more elaborate schemes to only retain basically the data on the network that's relevant. And to give way for people whose data is not retained to revive it, right? So for example, if we just retain the last year of data and you go away for five years, we want to give you a way to say, hey, here's how you can actually back up your data for the network. And so this big area is called state expiry or stateless depending on the approach.
“If we added rent tomorrow we would break every single contract. That would not be great. So we need to come up with more elaborate schemes to only retain basically the data on the network that's relevant. And to give way for people whose data is not retained to revive it...”
And to me, that's the biggest problem we need to solve after the merge, because basically, if we solve it, that puts a cap on how quickly can the network grow and how the demands to run a node keep growing over time? So yeah, others might have different opinions. Like, researchers working on sharding might tell you that their stuff is way more important. There's a lot arguments there too, but from my perspective, I think getting the merge done is priority number one. And once we get that, I think dealing with the state growth issue and finding a sustainable solution is probably the biggest problem we need to solve. And, you know, the merge plus that plus 1559, then we're busy for two years already. So we can have another podcast.
CR: Yeah, for sure. For sure. On this rent and like stateless Ethereum, it seems like a pretty controversial topic to me because, you know, when you're using a blockchain, you expect transactions to be there, the history to be there forever, you know? So it does seem like a controversial issue to implement something where you lose part of history. And a question on that, so you're saying the idea is that each user would be able to call up their own history, but for example, would a user be able to call up somebody else's history if they wanted to?
TB: Yes. So this might change, it's still active. But research is basically you keep two years of history a year. Like if you're in year three, right now, year two is the one that's still being edited. Year one is just kept as a snapshot. And then, you know, you skip over when you go to year four, you stop editing the previous one and whatnot. And if you want to bring your data back into history, you just need to provide a proof of your data and anybody can do that. So for example, if an application wanted to automatically revive the history of all its users, and maybe you could pay for that, right? Imagine you're using Uniswap and it is like a Uniswap premium or whatever.
Like you pay and it'll automatically revive all your liquidity positions and whatnot. They could automate all that. You don't have to be the person that stores it. But someone has to, and if you have part of history that like nobody stores ever, then yes, it would be gone. And that's the big downside. What I would like, which is a bit more complicated. But my kind of dream feature here is we first do that because it seems like we kind of know how you do it. And then I would like if we could add something like optional rent, where if you don't want to deal with having to backup your data or whatever, then you could maybe just pay the protocol and say if you don't pay rent, we'll clear you out after two years and you can back up however you want.
“You don't have to be the person that stores it. But someone has to, and if you have part of history that like nobody stores ever, then yes, it would be gone. And that's the big downside.”
Ethereum as a Nation-State
CR: That's a tricky one. Fascinating in any case, okay. We're coming up to the end of the of the hour. And this is one question that I really don't want to leave without asking you. Just on a broader level, looking at Ethereum in the long-term, what does success look like to you for Ethereum?
TB: Yeah, I kind of hinted at that earlier. I already see Ethereum as a sort of digital nation. So I think having more and more basically, businesses is the wrong word, but organizations that are built and deployed on Ethereum and that manage, not only large amounts of money, but just basically all types and having the equivalent of a functional economy on Ethereum I think would be really valuable. It gives you somewhere to coordinate in the world no matter where you are geographical. And I think that's really cool. It's also like a playground that we could use to test things kind of like a testnet for the real world economies where there's things like, for example, quadratic funding on Gitcoin that's been really popular. Like if we can test drive this on Ethereum and show that it works, you know, with like a million, 10 million then 100, then a billion then $10 billion. Then at that scale, it starts to be viable to bring those ideas back, like into non-blockchain economies and say look, hey, this has been running with $10 billion for five years. So that's really how I like to think about it. Yeah, it's sort of economic testnet for the real world.
“...having the equivalent of a functional economy on Ethereum I think would be really valuable. It gives you somewhere to coordinate in the world no matter where you are geographically. And I think that's really cool.”
CR: Nice. So, right. So when those testnets, or experiments prove to be successful, Ethereum can have an impact that starts to touch outside of its own ecosystem. What about Ethereum being the layer where like the real world happens?
TB: I mean I think for sure it's on its way to becoming there and it kind of has to, if it wants to gain credibility in that. But I guess luckily for people working on Ethereum, there is more and more economic activity that's becoming digital. Right. And I think Ethereum should probably try to capture as much of that as possible. But there's always going to be stuff, we still live in the physical world. And, and I think, it would be good to be able to export some of the innovations that we create. Like if you think of, again, Ethereum as a nation, nations have imports and exports and it's almost like we import capital both financial and talent from everywhere. And that helps Ethereum grow. But I think if you want it to be a lasting thing, thinking about what is going to be the export of Ethereum and I think that's its new economic ideas or models that we've tried in fairly contained environments that might be applied on a broader scale. So that's my hope.
“...if you want it to be a lasting thing, thinking about what is going to be the export of Ethereum and I think that's its new economic ideas or models that we've tried in fairly contained environments that might be applied on a broader scale.”
CR: I love it. I love that idea of Ethereum as a nation with its own imports and experts. Hopefully all of those economic experiments will have a positive impact in the outside world.
TB: Or if they don't at least we contain it to Ethereum. We don't have to break everything else.
CR: Just like, you know, keep yield farming, food coins in this limited environment. Well, Tim, this has been really fascinating, thanks for your patience and your time in explaining all these complicated concepts so clearly. I'm sure this will be super helpful for everyone listening to understand all the big changes and milestones that are coming up for Ethereum. So really appreciate it.Thank you so much for having me. This was super fun.
TB: Thank you so much for having me. This was super fun.
The Defiant is a daily newsletter focusing on decentralized finance, a new financial system that’s being built on top of open blockchains. The space is evolving at breakneck speed and revolutionizing tech and money. Spread the word and share!
