Mar 15, 2024

Tim Beiko from the Ethereum Foundation, shares the Importance of the Dencun Upgrade, what are their concerns, and Ethereum's long-term scaling strategy.

Video transcript

It's really hard to predict the future of the space. So I try to avoid it if I can. I think the trade offs for like the number of layer twos is, do you care more about like liquidity and network effects or do you care more about prioritized execution? Hello and welcome to the Protocol podcast. I'm Brad Cow here with my co-host Margot Neikirk and Sam Kessler. We're excited to dive into today's show with the latest news and developments in technology behind crypto and blockchains. First. Please do not forget to subscribe to our weekly newsletter, the Protocol on and let's dive right into it. We are very lucky today because we've got a big story this week which is good for journalists and we have exactly that right guest for this week's podcast. Uh The big story is the Ethereum D coon upgrade, which is the biggest upgrade for probably almost a year, at least um on the Ethereum Blockchain, which of course is the world's second largest Blockchain by market cap. But you know, also the first super influential smart contracts Blockchain that has really set the tone for kind of all of Blockchain development now that we're all seeing um even on Bitcoin and uh this upgrade is super important for Ethereum scaling roadmap. And our guest is Tim Biko uh from the Ethereum Foundation. He is the coordinator of the week biweekly all core devs call for Ethereum devs and he's joining us from Vancouver. Welcome, Tim. Thanks a lot for having me and Margo is our Ethereum beat reporter and she has been all over this reporter for not just like weeks, days, but also months and like a year. I remember when you wrote your uh D Kon preview sometime like probably last spring Margo maybe. Um Yeah, we wrote a Brangelina story about Den Kon about where that gets, gets the name from, right? The dub and the and the uh Cancun upgrade. So you know, these upgrade Tim can attest to this probably, but these upgrades love to Brangelina, the the two names because of the two upgrades simultaneously going on at the same time on the different layers. But um yeah, why don't we get into this? Um Let's start there. So we're talking about this upgrade uh Den Kon, but maybe Tim, you can, you know, just to set the stage for our audience like what are in, in your view, the key points, the key takeaways of what this d upgrade entails. Yeah, of course. Um The main thing, I mean, this is everything everyone talks about is uh E IP 4844 also known as Proto den Sharding. And the idea with that is Ethereum will introduce uh Ermal data storage to the network which we call blobs and layer twos uh which are kind of these, these constructions that Ethereum uses the scale in many different ways. They consume a lot of data on Ethereum and therefore they have to pay a lot of storage costs. And as a layer two user, if I'm sending a transaction on optimism, arbitra ZK sync, most of my transaction fee is actually paying for the storage cost of data back on Ethereum L one. And so with this upgrade, uh we're introducing ephemeral data storage and because the data is not stored forever, then obviously it ends up being cheaper and therefore it, you know, the those cost savings get passed on to layer two users. Uh So that's definitely the biggest part of it, there's a ton of other stuff in the upgrade. But um yeah, that's the focal point. I'm curious, you know, you, you brought up the the the focal point is for layer twos and roll ups, especially because they are sort of now part of this like bigger road map um and part of like the bigger Ethereum ecosystem and are sort of essential to scaling the Blockchain. And are there to sort of also allow for cheaper transactions? You know, from your perspective, there's been a lot of conversation and rivalry these, these layer two sort of who can deliver the best product and with this upgrade, you know, they're going to also be competing for lower transaction fees. Like, what do you sort of envision in terms of like that reduction in terms of transaction fees? Like, what is that ballpark? Um And do you have any sort of feeling in terms of what, how they aren't anticipating this new change, basically? Yeah, there's a lot to unpack here. So the, and, and I guess we're getting, we're getting into the weed. So the way the way the upgrade works is it basically segregates data for L twos in its own fee market relative to like all the other gas transactions on Ethereum. Um And this means that there's just like way less people competing for that. So like right now if I'm say sending a transaction on that one and sending you some tokens or doing a swap or whatever, um I'm competing for my transaction gas price with the optimism sequencer and the ZKC sequencer um on, on like this a option and therefore, you know, like uh it raises the price for everyone. So by having the separate free market, you allow only L two to like compete for this limited amount of blob space that we call it and that lowers the fees. It's really hard to say like exactly how much is this slow the fees? Is this like five X 10 X or whatnot? Because fees are a function of like supply and demand. And so um you know, like, sure you could say like at today's prices and usage and like the demand for, for block space, you know, if all those transactions moved, you probably see something more of like a 10 X reduction, but then if they all start using blobs, um and then more people want to use roll ups because they're cheaper. Obviously, this is so like prices going back up, but you're serving more people at that, you know, previous price point. Um So I I don't wanna make like a specific prediction on like, you know what the market equilibrium is. The way to think of it is like, you know, price is a function of supply and demand that we're increasing supply significantly for exactly the type of data that roll ups needs. And also the type of data that's not pretty useful for any other applications on L one. And the implication there is like, you know, this creates like much more room for them to post their data and and therefore lowers the fees. And the other thing that's really important about this update is inter has like much more long term scaling plans. So we call this proto bank sharding um because it's kind of the first step towards full sharding. And what we've done with this upgrade is we sort of set up all the infrastructure and scaffoldings for roll ups to use this and, and in the future when we kind of continue scaling and, and just have more basically blob space on the network roll ups won't need to do anything to support it. So they'll need to like change things now to move from like this permanent data to this, this temporary data. But in the future, as we come up with like different scaling mechanisms to add more of this temporary data roll ups will just sort of get those upgrades happening in the background without having to do anything. And I'm glad we're getting into the weeds. I also sort of jumped ahead of, of all this, but we'll get into that in a second with Sim and Brad. Um But I wanted to ask one more before I'll turn it over to them. We are speaking to you quite frequently. I feel like on a or on a yearly basis at this point in terms of like the upgrades itself, right last year, around this time, it was April, we had Chapela. The merge was obviously like the big, like the big upgrade that everyone sort of was looking forward to. I guess it's a year and a half from year to year. It seems like Ethereum sort of goes through like one major big upgrade where does um Dan KN rank in terms of like the most important upgrades to Ethereum compared to these past upgrades in terms of importance, a hard question in a way because um the obvious answer, I guess is right after the merge, um I think the merge will probably never do anything as significant and it's probably good as well. I don't want to do those things every year. Um But I think, yeah, probably right after the merge and the, I guess there's a couple of reasons for this one is um just the amount of like complexity and like sort of be risking that in a way knowing that you, that you can ship this this whole additional layer of, of data to Ethereum. And that's a pretty huge milestone. And then I think also the fact, like I I was just saying that it enables a lot of the future scaling is another super valuable property. And then the third bit would be um Justin Drake's been talking about this idea of phased roll ups uh a time in the past few months and based roll ups would basically leverage all of this as well. Um And I think this is another part we're really laying like valuable infrastructure with this upgrade, even though it's maybe not as like not all visible in the same way that like the merge was, you know, a, a very distinct shift from proof of earth or proof of state. And this time around you sort of see the tip of the iceberg with like, you know, the the small number of blogs we're adding. Um But most of the work was actually just like the infrastructure that to get this in, in a spot where we can build on it in the future. So, um this is the first time that um Ethereum has kind of taken a swing at sharding. This is proto den charting. It's kind of like this early version of the concept of kind of like adding lanes to the Ethereum highway, right, to improve, to increase capacity to, you know, increase the amount of track it can handle. In order to reduce fees this time, it's focused on roll ups. So we're gonna be reducing fees for ideally reducing fees for the roll ups that people use the layer twos on top of Ethereum rather than the base Ethereum chain, which is super expensive. So first off, just like orient ourselves is charting still on the Ethereum road map. Like are we still going to see fee reductions for the base Ethereum chain? That might be two questions, not just one. And then I've got some follow ups again, fee reduction is hard because the function of supply and demand. So, II I tend to think of it just like, can we increase the why of block space as like a given level of security? Like obviously, there's a bunch of ways you can increase the block space and make the chain less secure? We don't want to do that. Um So like, yeah, can we increase the block space significantly on L one? Um like producing security and the main blocker to doing that today is there's a lot of ways you can, basically, there's a lot of ways you can deal with the history of the state, you can deal with like all these, these like temporary temporary data for roll ups. But the biggest part of the theorem that we sort of can't reduce down is what we call the state size. So things on the theorem is everybody's balances all the contracts, all their, all their code, all their storage. So, you know, think like I own some. So there's somewhere in a theorem state that says, you know, my address has this amount of E and I have an ens. So there's like an ens contract somewhere that says, you know, Tim Mao dot E owns this address and there's a function by which I can buy and sell this token. Um So all those things, this is not even like a historical balance. It's just like the current view right now, we need an effort distort us. This is what we're sort of bound by. Like why couldn't we just have 10 times more transactions? On the theorem is the amount of new addresses, the amount of new contracts would just grow uh super quickly and then it would be really hard to run a note. So, so the way we, we think, you know, this could be approached is this idea of like stateless Ethereum. Um So today on the theorem, when I send the transaction, all I send is like the exact changes I want to do on this state. And the assumption is every node on the network has a local copy of that state and they're just gonna update it again. One way you could, you could change things is if every node only sort of part of the state or even none of it at all. And when I send you my transaction, you have like the state's root. So you know, like the what everything sort of hashes up to and then I send you like a part of the state I want to touch and what I wanna do on it. Um And this is something we'd like to do and we've been working for a really long time. The challenge with doing this today is that sending these parts of the state is really uh bandwidth heavy today. So the way the way through them stores the state is kind of like a Christmas tree. So it's like this long and narrow tree. And if you want to send a proof, you have to go from like the tip all the way to the bottom and you have to send this along with your transaction. So this is just like not tractable today. Ethereum has this other road map item which we call the verge, which is about changing how we store all of the state on Ethereum. So going from like a Christmas tree, something that's more like a bush which is like super wide but very narrow. Um, and then if I want to send you a proof, I need to send you like a shorter amount and we could gossip that. Um, so this is actually the thing we'll be working on, not in the next fork but the fork after that, we've already started working on it. But, uh, it's still gonna take some time so we can probably do another quick fork uh be before this goes live. But this is effectively, you know, I think the only way you can, you can raise uh the amount of L one block space without making any compromises is if you're able to provide proofs every time like a transaction is generated. And then if you want, if you want people to actually be able to send and verify those proofs, you need, you need them to be lightweight. So, so we have like this long road map of like, OK, we go from the Christmas tree to like the, the, the sort of bush and one of the things actually in this upgrade uh that enables that is deactivating the self destruct upload. So on Ethereum, uh we have this up code called self destruct which basically deletes a contract. And in our Christmas tree model, it's not the end of the world, you just like find the place in the Christmas tree and you, you sort of cut off that branch. But when we're gonna move to this like Bush model, if you wanted to delete a contract because it's spread out across the entire Bush, you have to search the whole thing and that's just not doable uh and like a 12 2nd block. So in this upgrade, the one that's activating tomorrow, we deactivating this off code which allows us to transition from one tree to another in a couple upgrades. And then once we have this transition, we're able to start building nodes which don't require a full copy of the state. And once we have that up and running, then you can raise the, the gas limit much more because you're, you're sort of uh not relying on everybody having a copy. Um It's a long way of saying like yes, we care about L one. but it's like a complicated problem. And in the meantime, if there's stuff we can do to help L two scaling, then that feels like a like a no brainer to do. Hey, Kim, you know, I think I I've read the, is it 18 days? Is that how long the the blobs are stored? Yeah. And, and why, why did you choose that number? Like I, I think maybe there's something there. Yeah. Um So that the main reason uh or the main constraint on that number, sorry is the roll up exit window uh for optimistic roll ups. So if you deposit money in an optimistic rollup, they credit it to you immediately. But if you withdraw your money, the way an optimistic rollup works is they post your withdrawal back to L one and they tell people you have seven days to dispute this and because they've posted all the data of all the transactions back on L one. If like I thought you had like a fraudulent transactions, I'm able to go to L one. Look at all this data, I'd be like, oh no, actually this withdrawal is invalid because look at this transaction, they make this thing that was illegal. And so I can like dispute that and block your withdrawal. So roll ups already have this assumption that if you've made your, if you've made your withdrawal on L one, you've waited seven days, you're sort of good to go. Um And that's how the L work today. Grown ups don't exactly have the same process. It's more a matter of like a couple hours than a couple days. Um But there's still this, this idea that like you want to, you want to verify the data that's been posted to L one and you want to be able to retrieve that data on L one. So 18 days is effectively, you know, like 23 X the like seven day window. So it's like a comfortable buffer around that. Um And also because this is like a new upgrade, um if something went like really, really wrong in the entire history of a theorem. I'm like, it's probably not anything we haven't fixed in like two weeks. Um So, like, imagine something goes terribly wrong and we needed the first bit of data before we had like a fix out or something. Um It would give us enough time to do that. Um But the, the real line constraint is just that roll up exit window and, and, you know, Margo obviously writes all the time about the race between all the L twos and, and there's the competing technologies of optimistic versus ZK. I'm curious, you know, in your opinion, is this like a technology neutral upgrade or is this favor one or the other technologies? I'd say it's pretty neutral. Um Yeah, like there's, there's probably a bunch of ways you can nit pick on both sides. Um But yeah, in general, it's like both, both roll ups can use the blobs. You know, there's no, like it's not like impossible for ZK A roll ups to use the blobs or optimistic uh to use the mobs like they can all use them. Um There's a bunch of work that, you know, they had to do to like, support them on both sides. Um Yeah, so I, yeah, I, I don't think it's significant. He favors one over the other and this update falls in line with folic roll up centric road map. The idea like you were talking about before of kind of fixing ethereum capacity problems not by necessarily reducing fees for transactions from the base chain. But now at least in the short term reducing fees for using these layer two roll ups that which are really the main platforms that people are using to transact on Ethereum today. But there's like a couple of issues with that that I wonder like if you can talk to us about uh that that members of the community have raised. So like II I mean, there, there's several things first, there, there's fears that just these roll ups might lead to fragmentation of the Ethereum ecosystem of, of liquidity of developers and so on to all these other disparate networks. And then there's also kind of the, the, the fear here that Ethereum by shifting to these roll ups is kind of increasing the community's appetite. Um and, and, and comfort with chains that don't hold the same sort of security decentralization principles. At least at this current moment as Ethereum does, they don't have the same security apparatus using them is not the same as using Ethereum, which is, you know, very secure, really robust and also has this decentralized network, but also has all the constraints with fees and so on that, that come with something like that. So there's a fear that by allowing people to move to these L twos or pushing them to the L twos, you're kind of making people more comfortable with a very anti sort of a vision. So I I that's kind of two questions again. But I wonder you know, how, how are you thinking about those, those sorts of issues with the roll up centric road map maybe just start with like the developer fragmentation and what not like I think this is like much more featured in a bug. So like different people being able to deploy different types of roll ups and experiment with that is hugely valuable if you are bottleneck on like Ethereum L one cord des for every like innovation and, and just like transaction computations. Um like it take forever. And I mean, like, look, we're shipping sort of uh the first part of, of sharding right now. Um If we had to ship a roll up as well, like we wouldn't have neither an optimistic nor a ZK roll up live in production. And now there's like many of both. Um So I think this is like, actually like that was actually a great decision in hindsight that like let the market come up with all these solutions and seeing, look if people, if people prefer to use optimistic roll ups because they can copy and paste their contracts from L one and braids if people prefer to use ZK roll ups with different V MS because they're able to use like a, a native VM that doesn't have the same constraints as the EVM does because it's operating in a different environment than they can do way more performance. Dxdydx was probably the first, like, you know, real good example of this. Um, that, that's amazing as well. And like, the fact that we have both today is something I'm, I'm pretty, I'm stoked about this. Like, um, I would not want like, the decision, like, do we do optimistic or ZK roll ups to have to, like, come through all four A and be debated for years. It's like, no, just start to ZK roll up and they'll see if people use it. Um And I think, you know, I feel similarly in terms of like liquidity, like, but you know, if people want to use a big roll up, like there'll be network effects there. Um If, if like moving liquidity from one roll up to another is like a big enough problem, then people will provide a service for that. Um So II I lose like absolutely zero sleep over any of this. Um I don't think like it's e the role to dictate how like the whole market structure should evolve. Um And I'm glad there's just more options for users. I think in terms of the security uh assumptions like that is valid, like roll ups today are definitely not as secure as the theorem L one for a bunch of reasons. Um And that's it, you know, they are moving towards that. And I think, you know, teams like L two beats, for example, do a great job of documenting exactly what is like the state of things and, and like what is like, how did they fall relative to each other? But also like, how do they actually implement things? Um you know, in detail. And so I think, yeah, it, it does fall on us as a community, like educate people around that. But the the flip side though is we can't prevent people from like going to other chains where it's cheaper, right? Like, you know, if the I'm like, look, we have an L one, it has these properties around security and this limited amount of block space, people bid to, to compete for this block space. Um Clearly like the price of that is higher than a less secure chain for a bunch of reasons. Um I think what we can do though is say like, OK, we have an idea for a construction that can be secure with roll ups. Um theoretically it sound but then there's always all these like implementation details to get there fully. Um and be like open about like, look, there is this thing there that has these properties, it's cheaper but like, you know, you're making these trade offs and ideally compared them all against the other and, and allowing users to make a decision. I I think that's like the best we can do. There's not a world where like Ethereum can say like don't use cheap chains, right? Like people will, we can at least try to find a way to get cheap and secure chains over the medium term. Fascinating. T thank you so much. Uh All right, we're gonna take a quick break and uh and we'll be right back calling all developers score a consensus 2024 developer pass for just $109 but act fast. Only a limited number of these passes are available. You may have heard that consensus ain't for devs, but here's why you're wrong. Consensus is the only place you can fully immerse yourself in a multi chain environment and learn directly from 20 plus chains, including arbitra chain link, Solana and more. Enjoy three days of intensive learning with technical talks, 40 plus expert speakers and 20 plus in depth workshops including dedicated half days for Ethereum and Bitcoin and three full days of programming on our protocol village stage. Consensus 2024 is happening. May 29th through 31st in Austin, Texas. Don't miss your chance to network at curated developer meet ups, find new career opportunities and explore hundreds of site events and hacker houses around town. Grab your $109 developer pass today. But remember, this exclusive offer is limited visit now to secure your developer pass before they're gone, explore the epicenter of Blockchain innovation at consensus 2024 and we're back with uh Tim Baco of a theory and foundation. Um Tim, this is a great, great conversation. Thank you so much. Uh OK. You know, I, I just say that sometimes I say that, you know, Bitcoin in, you know, it's 15 years of existence. 2009. Did I do that math? Right? Uh is, um you know, has already earned a chapter in monetary history. I think you could say the same thing about Ethereum, you know, I mean, just incredibly successful, incredibly influential and, you know, it's fascinating to see it all how it's all coordinate, it's all, you know. So these cod dev calls are also global. You know, people have talked about how Ethereum is the coordinations and the attempt decentralized war has maybe slowed down the pace of some of these innovations. We're already seeing some of these, you know, the, the problems with the congestion on Ethereum and the fees and we've had all these, you know, alternative data availability solutions that have come around and are now basically offering maybe I guess, kind of compete with the theory D A. But I'm just curious if you have any thoughts on kind of those bigger tradeoffs and what, you know, what that says about Ethereum, current state of development or adolescence or adulthood or, you know, title or, you know, where is, where is Ethereum and it's in its growth. Yeah. No, that's a great question. And maybe like the first thing I'll say is we absolutely take a sort of efficiency hit in order to maintain a more open and like permissionless system, right? Like there's no way, like if there, you shouldn't think of it as like a company or something like that. Like, obviously, if it was just, you know, 20 of us and you get to choose everything and when there's a disagreement, you get to say like now the CEO chose act and we're doing X um Yeah, we chip way faster but we chip like much lower quality stuff. And I think this is the thing that, that is kind of wild about Ethereum. Is it this open protocol um to a degree, you know, that like not a lot of, of, of chains have where the Ethereum specification is not even like production software. It's basically just more like an instruction manual, right? Like we agree first and forth. I was like, what are the rules of this thing? And then we have these 10 different teams that go and they each implement their version that takes a lot of time, right? But the flip side is you get all of these smart people across these 10 teams who are like extremely opinionated, who, you know, have like different perspectives and whatnot to like argue with each other to get to the best result. And there's just no world in which like one, you could get those 10 teams to work together on the same code space. Like, you know, smart people just disagree on too many things. They like all work together smoothly forever. And then two, it's also something where like, you know, we can't really gatekeep if you want to come tomorrow and start building an Ethereum clients. Like it's permissionless, you just go, you write the code and over time you get people to use it and you know, it is like an extremely valuable property where it's not like, say compared to something like, I don't know, like phone manufacturing, you're like, or, or hardware manufacturing. Um you don't have like this lock in where it's like, OK, you have to like be like an apple device and like, there's no way I could go and make like a tweet to a Mac book or something like that, right? So yeah, this does come out of the cost, it comes at the cost of like, you have to like make space for those debates to happen. You have more open debates and there's also less of a, of a forcing function of like making the decision, right? Like I don't get to choose what goes in the hard forks, right? I get to like, run the call where people over time, you know, come to consensus on it. But if I show up tomorrow and I'm like, you know, guys, like, we've been talking about this too long, we're just gonna do XY and Z like these people don't have to listen to me, they don't have to like, you know, another boss or anything like that. I, I think we've gotten better at like the processes of things. So like, you know, running our open process and like a more efficient way. Um And I'll say one, like big, big accelerant, there has been just uh investing in better testing and like infrastructure that all the different teams can come and use. Uh I think this is something in the past like 34 year since working on the merge where like the EF and other orgs have just like doubled down on like, can we build pretty good, like basic infrastructure so that when we agree to do something, we can actually like ship it pretty quickly and not spend all our cycles, you know, like trying to figure out how it's all gonna work together. But yeah, I, I don't know where that puts us in, in, in like a stage of development. I think everybody involved values the openness of the process, you know, over time we get better as like executing and, and just like working together with more and more people too. Like, I think this is the other big difference where today there's about 10 core client teams. And before it merged, you know, there were also, there's maybe eight, but they were separated in like four and four. We didn't really have to work together. So the merge, we sort of went from 4 to 8 having to work together and then we're, we're basically up to 10 now. And now we also have these teams like in this last upgrade. Uh Folks from Coinbase, folks from Unisoft, folks from optimism all like contributed um like serious engineering resources to make this happen. So um yeah, I one thing I'm super excited about it just it feels like we're able to do this fairly efficiently at scale in the open. That's pretty unique. I wanted to ask because um Brad had briefly mentioned about data availability solutions and you know, the the core of this upgrade is about the data the on Ethereum and that it's gotten more expensive. Um And as a result of that, like these D A solutions like Celestia like IND A like avail have sort of come to life to sort of solve that problem for now. Um But I'm wondering like in terms of, you know, longer term, like where does that sort of fit into the like bigger um Ethereum like roadmap or ecosystem as like this is trying to solve that problem? But data is still probably more expensive on Ethereum. But at the same time as well, I've heard from these L two teams that like the best data is the L one data. How do you sort of evaluate those tradeoffs between these D A solutions and L one data? Yeah, I'd say that probably the main thing you can think about if you're not like a network in which I am not, it's just which validator set is securing, you know, your data and therefore like which trust assumptions are you are relying on. So if you use Ethereum, then it's like the entire Ethereum validator set is the thing that secures the, the blobs, right? Whereas other, other D A providers have different validator sets. But so like, that's like the main trade off. And you can think of this almost as like using like a different L one from Ethereum, right? Like if you use another L one with a different validator set and you're relying on that data set for the security of your, your assets. Um And you know, like different validator sets have different properties. So I think Ethereum optimizes for like a highly secure validator sets, you know, through like a high, just like economic security level uh through like a decentralized validator sets and all of that um and different projects can make different tradeoffs there. I think the other aspect of like, you know, different D A solutions is you can obviously compete on like a technical and design that you choose. Um But I think, you know, to be fair, um Other, other teams might ship, you know, other innovations quicker than Ethereum that are like a another improvement. But Ethereum can also, you know, use, use those things and integrate them over time if we think that they're like safe and secure enough. Um So this is probably the trade off that's just like harder for an average user to see is just, you know, what is the actual technical implementation? Is it like a genuine improvement or is it just like making a compromise somewhere down the stack? Um And I think on the theorems uh in the theorems case, this is just not something we're willing to do. So we'd try to wait longer and it's taking longer to ship and potentially be more expensive in the meantime. But shipping something that we know, uh you know, we're highly confident in what it, when it feels like you mentioned this in the, in, in your last response. And the last question about like having the research and like the engineering resources from like a base and an arbitra and optimism sort of contribute to 4844. And, you know, as Ethereum sort of becomes not just like the core devs or like the core L one protocol, but sort of becomes this like whole other industry of roll ups and D A solutions and like all retaking, you know, you know, you're at the EF and that's like, that's sort of at the core of coordinating what's happening on Ethereum. How, how do you sort of see your role and how does the EF sort of play a role in coordinating with these other teams? Like, like you said, like ar like representatives from arbitra and base have like come on to join the, like a call about 4844. But I imagine like there's gonna be so many more upgrades that affect not only like the L one but also all these other mini ecosystems. What is the ef thinking about in terms of like getting everyone on the same page as this umbrella and this big tent gets like becomes an even bigger big tent. Yeah, that's a good question. I think there's like probably three ways you can break this down. Uh First is like coordinating the L two ecosystems or diversity ecosystem. Like I don't think the EF should be doing that at all, right? Like, you know, Arbitron can coordinate its ecosystem. A can coordinate its ecosystem like we don't, you know, have or necessarily want to have any say in that. Like, I think again, again, it's just great to have like different ecosystem behave differently and adopt different things. So um that's all up to them. I think the second part though is like, how do you bring in their voices when stuff affects them? Um So I think, you know, if we are doing something that's gonna affect uh L two s that's gonna affect uh rest staking, that's gonna affect, you know, even just like smart contracts general. But for example, if uh I mentioned we're removing self destruct in this, in this fork. Um When we did this, you know, me and others at the ef spent time like one, we just hired auditors to tell us which contracts would potentially break and, you know, find out if any and then reach out to them. But then we're like, OK, the L twos use self destruct in a way that's like, weird and different from L ones. Like, can we get them to actually sanity check this or like, are there other like big contracts? And I think like, yeah, I, because you know, we and the coordinating a lot of this stuff like, like how we coordinate the cross, like all the rollouts or all stuff like that. I think this has been like one of the hardest questions because we definitely don't want the ef to be a bottleneck to that. And it's like finding the sweet spot between like, can you, you help and like help them all, like talk to each other but not have them depend on you uh or like, you know, ask for your permission or something like that. So we've, we've launched an initiative that the trial that is called roll call where we, we try and discuss like potential standardization stuff for L two S. Um And I'd say it's been, you know, it's, it's been, it's been all right so far. Um A couple of things we've had is like L twos use different pre compile which are like special contracts on, on Ethereum. There's like a range of 1000 addresses for these contracts. We, we used about like 10 of them give or take. So we agreed, you know, based on these discussions about the L twos, they want to do a bunch of more special contracts. So we agreed, ok, we're gonna give them like half the space, ok? We have 1000 1000 addresses reserved and you can know the flight from 500 to 1000 L one is never gonna deploy anything there. Um So I think that's like a small example of like, you know, it's not like a huge amount of, of change, but by L one team saying like, you know, this space is yours, it sort of creates more certainty for the L two to go and debate. OK. What do we actually want to do in those 500 addresses? L one doesn't have to have an opinion about that, you know, you can sort of do whatever you like. Um But like at least, you know, OK, this is like the, the sort of sandbox where like, um you can go and run these experiments. So I think for things like that, like it makes sense for, for, you know, the EF and for like just L one coordinations to sort of extend there, but I would never want to see it go to a level where it's like, OK, you have these 500 addresses and we think you should do, you know Xy and Z and you know, those are the good ones and you know, these other ones are bad ones like L two S should, should be able to, to govern that as we're starting to kind of wrap up here. I wonder if we can kind of look past um Den KN towards the next step on Ethereum road map. Um The next set of updates, Prague and Electra, can you tell us briefly like what is going to be in those packages? Um And what should we look forward to over the next year or two of theorems development? Yeah. So Prague Prague and Electra um is is going to be a relatively small upgrade. So since the me, we've sort of got into this pattern of doing like one really big fork and then a smaller one and then this is the bigger one coming up and then uh Prague will be a smaller one. And the reason for that is it allows us to work on like two forks in parallel. So I mentioned all the stuff earlier um viral trees in that transition, we are working on that, but it's gonna take, you know, more than a year. So in the meantime, we have bandwidth to like work on other small small wins and similarly on the consensus layer, uh they wanna keep scaling Ethereum providing more, more blob space basically. And there's a bunch of work on that, but it's also gonna a while. So the idea with Petra is to try to find like a bunch of small wins that we can get relatively quickly while we, we sort of prototype the larger stuff. And so far we have things like triggering, uh triggering validated withdrawals from smart contracts, which is like a very nice thing um that you can't do today adding the BLS pre comp pled, which is the curve that's used on the beacon chain that you can't verify natively on L one debate. Um and or, and removing uh the deposit window. So when you, when you make a deposit for a new validator, you have to wait 2000 blocks for it to be activated. This was in the proof of work days. Uh Because, you know, we didn't want to have reorgs past that point, obviously, since the merge, we, we don't need that anymore. The validators know every block sort of what's up on the, on the execution chain. Uh So we're gonna remove that in the next fork. Um So you can think of it as like just all these sort of small but valuable things and there's a big conversation around wallets and Eoas and account abstraction as well. So we might see a fork uh or sorry, we might see an E IP in that fork that, that addresses that somehow. Um So you can think of it like, yeah, a bunch of small nice things coming in Petra, hopefully some time this year um early next year or something like that. And then in parallel, I often worked done on like these larger initiatives um on both the execution and consensus site. Yeah, I mean, one of the great, you said we're going into the weeds. I think, uh one of the great things is, is having somebody who really knows it, explain it because you, you uh explain it in very simple terms, which is very much appreciated. Uh And ok, just one last one here. Um You know, I was talking with our editor in chief, uh Kevin Reynolds just a little bit ago and we were talking about the number of layer twos and I just pulled up the website L two B which I'm sure you check occasionally, maybe you do. Um And they list 45 active layer twos, 34 upcoming projects. And then they've, they now have a layer 33 projects, they got four active. But, you know, I'm just curious extrapolating out. I mean, we included something in the protocol newsletter that, you know, one person was predicting 10,000 layer twos by the end of this year. But I'm just curious, you know, as this goes through, you know, what sort of, you know, what's it gonna unleash in, in your view? Yeah, it's, it's, it's really hard to predict the future of the space. So I try to avoid it if I can. Um I think the trade offs for like the number of layer twos is, do you care more about like liquidity and network effects or do you care more about um prioritize execution? And what I mean by that is if you have like a specific app that everybody needs to like touch at the same time. Um And it's really important that like you touch it all the time, like you like gaming is usually the example that people come up with this, right? Like you need to always be like touching that contract over and over. It kind of makes sense to have like your your own chain because you don't want like other transaction. So like uh slow down the rate that's which you can, you can post your updates. Whereas if you have something that's like super interconnected, right? Like defi is a great example of that where I want to like, you know, put some collateral borrow against it, take my Boral collateral, put it in some other protocol, then you care more about like the network effect and like getting say those three transactions done in one block. So I wouldn't be surprised that we see something like kind of two types of layer twos. One is like just like effectively similar to L one where you have, you know, like a big interconnected ecosystem and then you see like way more specialized chains where people want to go there, they want to do like this one thing super to be super specialized. Um and maybe it's not just one, right? Maybe it's a handful, but they, they have like a more narrow use case and then they optimize everything around that. And then once you're done with that use case, you know, you can bridge back out from like that specialized L two to like a more generalized one. So um medium confidence in this again, predicting the future is hard. Um But I think it, it will be possible to have like many, many different types of L twos. And I mean, it's already the case today and with the blobs as well, I I noticed a bunch of people already working on like um sharing blob services. So when you, when you set a transaction, will it, you need to, you need to fill up the full blob. But then if you have a bunch of small L two S maybe they don't. So people are working on like a service would just say, OK, you use 10% of the blob, you use 20% and then, you know, we could put like five different L two S in a single blob. Um So I'm sure all of that will emerge and we'll see about four L two S I, I wanna ask like 10 questions about what you just said. But uh I think we had a wrap and, and I know you've got a busy, uh you're quite busy yourself. So uh Margot, thank you for leading this uh fascinating discussion. Tim, thank you so much.

Learn more about Consensus 2024, CoinDesk’s longest-running and most influential event that brings together all sides of crypto, blockchain and Web3. Head to to register and buy your pass now.