Bankless - The ZK Fast Track | Andrew & Uma of Conduit & Succinct

Episode Date: December 4, 2024

Last month, Andrew Huang (Conduit) and Uma Roy (Succinct) announced OP Succinct, a groundbreaking collaboration that blends Succinct’s SP1 zkVM with Conduit’s Rollup-as-a-Service platform. This new zkEVM promises 1-hour finality, breaking away from the traditional 7-day delays seen with optimistic rollups. In this episode, we unpack what makes OP Succinct a game-changer and how it challenges the conventional ZK rollup landscape. Why is this Rust-based zkVM approach so different? Is ZK rollup tech finally ready for prime time? And could fast finality and the end of fraud proofs be the key to solving Ethereum’s fragmentation problem? ------ 📣SPOTIFY PREMIUM RSS FEED | USE CODE: SPOTIFY24  https://bankless.cc/spotify-premium   ------ BANKLESS SPONSOR TOOLS: 🐙KRAKEN | MOST-TRUSTED CRYPTO EXCHANGE https://k.xyz/bankless-pod-q2    ⁠ 🦄UNISWAP | BROWSER EXTENSION https://bankless.cc/uniswap  🐧CARTESI | LINUX-POWERED ROLLUPS https://bankless.cc/CartesiSimple  🛞MANTLE | MODULAR LAYER 2 NETWORK https://bankless.cc/Mantle     ⁠  📈iYield: YOUR FINANCIAL PICTURE, SIMPLIFIED https://go.iyield.com/bankless    ------ ✨ Mint the episode on Zora ✨ https://zora.co/collect/zora:0x0c294913a7596b427add7dcbd6d7bbfc7338d53f/106?referrer=0x077Fe9e96Aa9b20Bd36F1C6290f54F8717C5674E  ------ TIMESTAMPS 0:00 Intro 2:52 zk Rollups Past vs. OP Succint  5:48 OP to zk Rollups Tradeoffs & Timeline 12:39 Benefits of 1-Hour Challenge Period  15:34 Scaling the Rollups  18:12 Endgame Vision  20:47 Cost of Proving Must Go Down? 21:55 Conduit’s Distribution Role 25:55 AggLayer vs. zk & Adoption  29:57 zk Rollup on ETH L1 Constraints? 35:53 Ethereum’s Negative Sentiment 51:04  Closing & Disclaimers  ------ RESOURCES Uma Roy https://x.com/pumatheuma  Succinct https://www.succinct.xyz/  OP Succinct https://github.com/succinctlabs/op-succinct  Adrew Huang https://x.com/KAndrewHuang  Conduit https://x.com/conduitxyz  ------ Not financial or tax advice. See our investment disclosures here: https://www.bankless.com/disclosures 

Really interesting thing that we've seen with like Sysceton, ZK in general, is just the cost curves are coming down. And so if we can find a foothold today and kind of a low throughput chains or even higher throughput chains where users are willing to pay that additional transaction cost, it just gets better over time. And welcome to Bankless, where today we are exploring the frontier of ZK technology. It's been said that ZK is the end game. And it's also been noted that ZK technology has been progressing a little bit faster than people previously anticipated. Meanwhile, in the old world of optimistic roll-ups, we are still kind of stuck in stage zero, stage one of the roll-up decentralization. Our two guests today, Uma Roy from

Sucinct and Andrew Huang from Conduit, I think that they have a silver bullet to leapfrog a bunch of the optimistic roll-up stacks. What do we get with what they are delivering? The O.P. Sucinct is potentially hyper-scale roll-ups that are at-stallel. stage one or stage two that have what are like restful APIs between rollups to be able to talk to each other and fix some of Ethereum's fragmentation. Is this the silver bullet that they promise? Well, this is the big question that we ask in this episode and how does this fix Ethereum's fragmentation? I'm pretty excited about this. The fact that this is already out in production is pretty cool. So we will let you Bankless Nation judge for yourself about what you

think. But I love this episode quite bullish. All that sounds super brainy, David. But my biggest question owner of this episode is, does this help my eth bags at all? Okay. And that's a question. Could ZK. Supercharge the Ethereum roll-up-centric roadmap? Could it even supercharge the initiative to make L1 better? And those were some of the questions we asked at the end of this episode. In addition to this narrative funk, this sentiment that has invaded Ethereum of like, you know, prices down, it feels bad. What do they think about that? And so we weigh into this as to whether we're on the right track or not with the larger Ethereum roadmap and how ZK can help. I think there's been some consternation about the slowness of Ethereum core devs, even though the Ethereum roadmap is competing

with, you know, Celestia, Salana, Bitcoin on different parts of execution, data availability consensus, even though the role of the developers are all kind of going in parallel in three different directions, I think what people are forgetting about is that what's also happening in parallel is roll-up technology from third parties, from private interests, innovating and scaling the rest of the roadmap, the rest of the roll-ups to super, like I said, supercharged them kind of to their logical conclusion. So all of that is happening at once, and this episode is focusing on that part of the Ethereum roadmap. So let's go ahead and get right into the episode with Uma from Sysink and Andrew from Conduit. But first, a message to talk about some of these fantastic sponsors that make the show possible, especially Cracken, our preferred crypto exchange for 2024. If you do not have an account with Cracken, consider clicking the link in the show notes to getting

started with Cracken today. If you want a crypto trading experience backed by world-class security and award-winning support teams, then head over to Cracken. One of the longest-standing and most secure crypto platforms in the world. Cracken is on a journey to build a more accessible, inclusive, and fair financial system, making it simple and secure for everyone, everywhere to trade crypto. Cracken's intuitive trading tools are designed to grow with you, empowering you to make your first or your hundredth trade in just a few clicks.

And there's an award-winning. client support team available 24-7 to help you along the way, along with a whole range of educational guides, articles, and videos. With products and features like Cracken Pro and Cracken NFT Marketplace and a seamless app to bring it all together, it's really the perfect place to get your complete crypto experience. So check out the simple, secure, and powerful way for everyone to trade crypto, whether you're a complete beginner or a seasoned pro. Go to crackin.com slash bank lists to see what crypto can be. Not investment advice, crypto trading involves risk of loss. Are you ready to take control over your entire financial life, crypto, defy, and fiat all in one place. Meet Iyield,

the free financial planning tool built for crypto natives. Unlike traditional portfolio trackers that just show you the value of your assets, eye yield goes deeper. It consolidates everything, assets, debts, income, and expenses, offering you a complete financial picture across 16,000 tokens, 40 defy-foy protocols, and all fiat currencies. With real-time defyield tracking for platforms like Ave, Athena, Eigenlare, and Uniswap, Iyield ensures that you're always on top of your staking rewards, investments, and cash flow. flow. And the best part? It's 100% free with no ads and it's secure. No personal data, no ID requirements, no compromises. So head over to iYield.com. So sign up today and get the guesswork

out of financial planning. The Uniswap wallet is officially the preferred wallet of bankless. And it's the one we use any time when we want to transact on chain. Whether you're on your browser or on the go, uniswap wallet makes it easier than ever to swap anytime, anywhere. Use your wallet to transfer funds directly from a top centralized exchange and taping thousands of tokens across Ethereum. And over 10 other chains like base, Arbitrum, And optimism. Uniswap wallet delivers deep liquidity, fast execution, and reliable quotes with zero gas swaps through Uniswap X. And when it comes through security, you can rest easy knowing it's backed by Uniswap Labs, one of the most trusted teams in Defy. Their code is

open source and independently reviewed so you know it's protected. So why wait? Download the Uniswap wallet today on Chrome, iOS, and Android. And don't forget to claim your freeunit.orgia.com Start swapping smarter with Uniswap. Okay, so Andrew, Uma, last month, you two announced a collaboration called OP S succinct. And I'll read from the introduction of the blog post. We partnered with Conduit, this is from the Sucinct blog, we partnered with Conduit, a leading roll-up platform to support easy development of OP-Sync roll-ups. OP-Sync is a best-of-all-worlds Z-K roll-up. It combines the innovation of SP-1, our general purpose ZKVM with the OP stack to create a fast, affordable, and highly customizable

ZK-EVM. This is the first ZK roll-up stack that Conduit supports, and it is a new is the only one offering fast finality of one hour instead of seven days. So, guys, whatever is going on here is kind of breaking the mold for how I've traditionally understood ZK roll-ups to come into the market. ZK Sync, Polygon, scroll, these teams are all building their ZK framework stack as like their own flavor. Yet here we are with a Rust-based ZKVM provider combining with a roll-up as a service provider to bring a new ZK stack framework to the table.

Can you guys tell me what the hell is going to? going on here. Can you guys elaborate on this like alternative approach to entering the market? Umah, can I start with you? Yeah. So I think you had a great point, which is how were ZK roll-ups done in the past is really different than what we're doing with OPs has synced. And that's because it used to be the case that ZK was really hard. So you'd have all these teams like scroll, Polygon, ZK sync that have teams of 30 cryptography PhDs that have to build all these ZK circuits, which is basically magic moon math and make a really complicated thing, a very complicated machine to build a ZK roll-up. And there's a lot of downsides to this. Like the roll-ups are not fully EVM-compatible,

which means it makes life worse for developers. They take a really long time to develop. They're hard to upgrade. And it's all very complicated. And I think that's part of the reason why these roll-up as-a-service companies, like Conduit, haven't really been able to offer these types of roll-ups to the broader ecosystem. So that was the old way of doing ZK. Now, with SP1, ZK is really easy. You don't have to have 30 cryptography PhDs. You can just write normal code and use it. So ZK basically turns into a weekend project, and that's pretty awesome. And then what this lets us do is it lets us take an existing roll-up stack, Lego P stack, which is used by base, Unichane, Roll-Chane, and a lot of other teams. I think, Andrew, you guys have like hundreds of teams experimenting with it.

And you can take that and you can just take normal code for it, put it in SP1 and make it a ZK roll-up. So that's why I like to say it's kind of the best of all worlds because you kind of get the existing roll-up stack that's used by many teams. It's battle-tested. It's production ready. And you can get all the benefits of ZK, which is fast finality, better interop, you actually secure stuff like that and get the best of all worlds.

Okay. So are we saying effectively that like all of the optimistic roll-ups that we have, have, at least in the OP stack, like, it could be a weekend project to just upgrade those to ZK roll-ups? Yeah. Wait, for real? Seems a little too good. There's no trade-offs there?

We're used to the word trade-offs. Is there a trade-off in that transition? I think probably the only trade-off is that it's a little bit more expensive. Like, you have to pay for ZK-proving. Yeah. Or it's a little more expensive than if you don't have to do Z-K-proofs. Why this feels like a big deal is like I've long, I mean, we've watched kind of the dual development of optimistic rollups and ZK rollups all the while the entire industry's knowledge

like that ZK rollups are going to be the end game. So it's always felt like there's some technical debt out there with all the optimistic rollups will eventually have to kind of like become ZK rollups. And I was just under the naive impression that that would be like a really massive like roadmap item that could take like months to years to actually accomplish. plus the actual willpower and we might like forever be stuck on optimistic rollups. But this is just like if it's as simple as kind of some lines of code and flipping a switch, like is why do you essentially predict that all of the optimistic rollups will just turn into ZK rollups like sometime next year? Is there anything preventing that or could that even like why is that a good thing?

So we think the ZK technology is like incredibly impressive. I'd say it looks like it's happening over. a night, but Uma has been, like, building for, like, years. And, you know, I met her two years ago, actually at the Zerox Park kind of residency in New York. And back then, like, the state of the art of ZK was you're using Surcom, you're writing, like, these custom circuits that take forever to write. And there's no hope of getting these really complicated, like, EVM implementations to prove in a way that was computationally efficient and, like, not super, super expensive. I think the core innovations that have happened over the last two years are really one much better rust

tooling for the EVM so like i think georgios with uh you know i think ref and like our EVM opi alloy all this stuff has made developing the EVM in rust much much easier and then i think the core kind of innovation here is really SP1 which enables you to turn any kind of rust code into like a ZK AVM. And so that's really the composition that's happening here, enabling us to, in some ways, like leapfrog, a lot of the ZK implementations that are out there today. I think the main barrier to entry today is really, one, like the integration. So obviously, you know, weekend project to, like, get it, quote, working. It's a little more effort to get this integrated into a production-ready kind of framework. And we've been working with the team on that and we're excited for it. I think the

other thing that kind of comes up is really, and Uma mentioned, is kind of the cost. And so you might use this for a high throughput roll-up, and the way to make it scalable is actually to pass on the cost to the user. And Uma will actually have better stats here on the cost per transaction, but it's somewhere between 0.1 cent to a cent more per transaction, which if you look at most layer twos, a lot of users are kind of willing to pay. I think where ZK might have more trouble today is where, you know, teams. are subsidizing fees for their users, in which case it might be too expensive. But I think the

really interesting thing that we've seen with like Sysington, ZK in general, is just the cost curves are coming down. And so if we can find a foothold today and kind of a low throughput chains or even higher throughput chains where users are willing to pay that additional transaction cost, it just gets better over time. And really, I think the story here is if you look at optimistic fault proofs, particularly for large systems like Arbitrum 1 or Base or OPEM.O.P.M. that they require like huge bonds and these large kind of challenge periods and withdrawal periods because they have this like interactive kind of proof system. And so, you know, it's really cool technology. It's like big brain. It's awesome. But probably only the largest chains are going to be

able to afford the bond that secures that TVL. I think the really interesting thing about ZK is democratizing that security and that proof system to any chain that kind of wants it. And so I think That's the core innovation here in addition to a bunch of other kind of secondary effects like faster finality. There's two things, Andrew, that you said that I want to make sure I understood. You're talking about which of these, like, layered two chains are, like, the most appropriate for this upgrade. And I think what you alluded to is it's the ones that have a critical mass of users because the last I checked in on, like, my ZK knowledge, ZK really works when you can amortize costs across many, many, many transactions. And so if you are like a brand new layer two with like not a lot of activity and you're just kind of bootstrapping up your activity, the proving costs are going to be high per transaction.

But like maybe if you're base or arbitram, the big ones where you can really spread out the costs of proving across thousands and thousands of user transactions, this is when leveraging SP1 really makes the most sense. You guys are both nodding. So I'm sounding I'm getting that part right. I think definitely directionally right in that the more. So there's an overhead per block for pretext. but the more transactions you have, the more you can amortize. What I'll say that's, you know, what's been impressive is actually bringing down the fixed cost that you're referring to.

And so, like, a new chain can actually get started as easily as, like, three K a month for kind of succinct proving, which is, like, on par actually with the cost of, like, launching the roll up on conduit. And then the kind of volume-based cost or the usage-based cost, you can end up passing on to the user to the tune of 0.1 to one cent per transaction. And so there are a lot of use cases that, you know, particularly on the lower throughput side, that can support that, even on the higher volume side, I think if you look at priority fees and base fees on base, users are typically paying at least that much or even more than that. And so I think it's getting to the point of affordability where really anybody can use it. I think the main use cases where it's harder to do is where kind of the chain operators subsidizing transactions.

Because at that point, you know, you're already dealing with, you know, spam prevention and things like that. and ZK might be too expensive, and that's where ZK fault-proof might make more sense. Again, an improvement over the OP fault-proof because it does require a challenge, but it's non-interactive. And so you don't need a super expensive bond. And really, I think what we're seeing is like ZK almost leapfrogging kind of the optimistic technology. And this is such an underrated, like, kind of advancement on the tech tree that honestly we didn't see coming. And so when Huma came to like pitch us on it, like, we were. were kind of blown away and we're super excited by it. Okay, that was one of the things that I wanted

to get clarity on. The other one was the bond. You talked about the very large arbitram bond that arbitram has to use to facilitate withdrawal request, right? Like, it needs to have liquidity in the bridge to unlock assets that people are depositing and getting out of arbitram. And because of the seven-day optimistic fraud-proof window, they need to have a sufficient amount of liquidity to account for seven days of liquidity. And so, you know, anything can happen in seven days. But because, Because this ZK OP S succinct only has a one-hour challenge period, all of a sudden we're unlocking a lot of capital. Is that the main benefit of having a one-hour challenge period?

If we can just like, you know, unleash a lot of capital here and kind of restore that into circulation? Or what other benefits come with that one-hour challenge period? So the bond is separate from actually the liquidity in the bridge, and the bond is there and needs to be proportional to the amount of capital at risk, right? And so for Arbitrum, you know, there's like billions of TPL on there. And that's why their bond is like, you know, like millions of dollars. You might see the same thing on OP Maynett or B.

And so that's why it gets so expensive. I think what's interesting about ZK is because you don't have this challenge kind of mechanism, you're actually proving the execution. And the state route that gets posted to later one can only be accepted if the proof is correct. It means you don't need a challenge gain. And so with that, that means you don't need the bond. and you can have much faster finality guarantees. What faster finality actually enables

is much actually cheaper kind of bridging or interop between different layer twos. And so I think you're exactly kind of, you hit the nail on the head where if you have a seven-day withdrawal period, you know, fast bridges, whether it's like, you know, across, layer zero,

kind of relay, you name it. They have to take kind of the capital cost of, you know, shifting capital across these layer twos, and having this capital locked in these seven-day bridges, that gets, like, very expensive over time. When you bring that from seven days to one hour, that's just, like, an over 100x improvement

in the capital efficiency of these protocols. And if bridging is even, like, relatively inexpensive today, it goes even cheaper. And I think that's the type of interop and kind of primitive that I think enables cross-chain interactions in a way that we see today at kind of a minimal level in terms of, kind of bridging, but I think intense and interop,

is something that is going to be turbo boosted by kind of fast finality. Well, one thing I'm saying in the interop space is that it's increasingly finding ways to be abstracted and kind of pushed into the background. And if bridging costs also go down significantly because the cost of capital is going down significantly, that's just another way that we can just kind of like hide the costs of bridges. If bridge cost goes from 10 cents to a 10th of a cent, like you can kind of just like sneak that into just like a fee somewhere and hide that.

So that's pretty interesting. I want to really tap into what I think is kind of the main course of this upgrade, which is scale, scale, scale, scale. Uma, do you have any, like, kind of numbers that you can give us about, like, what you're doing to the roll-up space or what you can do to the roll-up space when we have SP1 integrated into the OPE stack? Yeah, so I think Andrew from Conduit is kind of of the opinion, which I agree, where we're going to have thousands of roll-ups on Ethereum.

And that's really how we're going to reach the massive TPS, the massive scale, that we need. And we already see that today where basically every big application or company like base, Unisaw, WorldCoin, they're all launching their own chains. I think something you said earlier is that today with all these different roll-ups, it's really hard to make them interoperate. Like bridging is a huge topic in the ecosystem, interoperability, fragmentation is like a lot of what's in the Ethereum narrative right now. And it's like negative narratively. And I think ZK really helps solve that. Basically, if you have every roll-up being a ZK roll-up, you don't have this seven-day nonsense anymore. And you just have all the roll-ups talking to each other in, like, an hour.

They can verify each other ZK proofs. You get interoperability within an hour in the worst case. And then in the best case, you have teams like across or layer zero who can just fast bridge. And basically, the way it works is they issue kind of like a, they kind of front the capital to the user, and then they settle in the background. Now, if you talk to Hart from across today, he will say his life is really difficult to set up all this infrastructure because he has to manage all these capital for seven days across all the roll-ups. So that really sucks for him and his team. And it sucks for any team that's like trying to set up these vast bridges or, you know, chain abstraction layers or whatever. But with ZK, like that seven days becomes one hour, which is literally over 100x improvement.

And it makes all those teams lives easier. So the biggest thing I'm excited about is kind of. to fix all the problems with Ethereum of fragmentation and interop, we're going to need chain abstraction. We're going to need all these bridges. We're going to need all these liquidity layers to make all the rollups feel like one chain. And I think that'll happen for sure. And I think the way that's going to happen is that it'll be a thousand X easier for all these teams to kind of manage their capital and manage all the bridging and liquidity and money across all the different rollups when all the rollups can settle and talk to each other in one hour instead of seven days.

So maybe to summarize, I think actually having fast finality and ZK level finality is one of the biggest things that's holding back the ethereum ecosystem today from feeling like one unified chain. Okay, this is amazing. This is why we wanted to have both you guys on is because I thought this was, and I think David thought this was like a distant future, right? And again, it's here and now. So like, I'm kind of asking what the, like, I mean, is there bad news here? Like, this seems like such a no-brainer for all of the. optimistic roll-up change to just upgrade at some point, right? Like, why wouldn't they all do this?

So you're telling me that we get better interoperability, one-year, like one-hour interoperability, and, like, we don't have that withdrawal period. We get to not worry about all the complicated fraud-proof code. And by the way, fraud-proofs were kind of a cluge anyway, right? It's just like it felt like at the entire times, like, you know, no one wanted to say optimistic roll-ups were just, like, using some clugee kind of, like, you know, better to replace that with cryptography whenever we can. So we have ZK roll-ups and then capital efficiency improvements. Anyway, at the end of 2025, is there any reason that all existing optimistic

roll-ups don't just upgrade? I've talked to the Arbitrum team and they have seemed in the past to be a bit less bullish on ZK than like other teams. So I don't know if they'd be a hold-up and a holdout. And if so, if so, why would they be like, is there any downside? to this. I think if you talk to the optimism team or the arbitram team, they actually agree with you and they think ZK's the end game. I think probably the biggest disagreement that me and Andrew might have with them is the time scale. I think the optimism team and the arbitram team thought maybe all their roll-ups will be ZK in like three to five years. I kind of agree with you, Ryan, that it should be the end of next year. I think the tech is finally ready. It's actually like cheap and

fast with SP1. And it's also really easy. Like the developer experience, a customizing, ability, the upgrade ability is really, really good with SP1. One thing to say is this tech is pretty new. Like, SP1 only went live on Mainnet in August. And, you know, we only got Opie Sistinct working. We demoed it in like late August, early September. And then we only recently got stuff running in production. And there's like a long cycle to take something from a really promising proof of concept of production. We worked a lot with the Condo team on that, even like late October earlier this month. So I think as with any big tech advancement, it like takes some time for these big companies like optimism and arbitram to kind of understand it, be like, oh, it's

actually real, oh, it's actually running. Oh, okay, let's use it. They don't want to be first out of the gate, right? Okay. But does this do anything for, you know, the L2B stage progressions in terms of like the level of decentralization of your roll-up? Does incorporating, like, converting your optimistic roll-up into a ZK roll-up, does it? Does it do anything to advance that cause? Can you get to stage one faster or stage two faster as a result? Yeah, today, most roll-ups that run are stage zero, actually. So most roll-ups deployed on the roll-up as-a-service platforms are not actually using the fault-proof technology

because it is pretty complicated to set up. It requires large bonds. You know, who's running the challengers? I mean, there's a lot of questions around that. And I actually think there's this kind of myth that fault-proofs. are simpler than ZK. I actually think they're a lot more complicated. You're totally right to call it this like Kluji technology. And there's a lot of hidden costs associated with what's going on there, including the bond period, capital inefficiency, stuff like that. So yeah, I think basically

like all those rollups today are at stage zero. When they adopt Opie Sinct, I think they'll be at stage one, which is a dramatic improvement versus just being a multi-sig. So one, guys, one part of the Ethereum roadmap, many of the EASC, many of the EASC, developers, Justin Drake, will say, like, it's left up to the layer two's. And that's really a lot of, like, the order of magnitude of scale and reducing latency. And maybe we can kind of fill in that part of the conversation of how this model, the ZK model kind of just completes the rest of the picture of Ethereum's like layer two scaling roadmap.

You know, Uma, you said ZK is the end game. Um, I think when Uma talks, we're talking about like, okay, here's one individual chain and we're going to just like juice that thing up to the max. And then we're going to turn to Andrew and be like, and then I'm going to do that for 10,000 more chains. So maybe, Uma, we can kind of go back to that part? Like, again, like, can you articulate just the narrative, the vision for how you can take this and scale a chain to its like logical conclusion? Yeah, so a natural question is if I have a really, really high TPS chain, like how does ZK kind of work? Like, does it become really, really expensive?

And the answer is actually no, because if you have a really high TPS chain, there's some. fixed costs for even generating the proofs at all. And those are actually really cheap, as Andrew was saying. It's like a couple thousand bucks a month. And that number is going down, like, day over day. And then there's like kind of each transaction has a proving cost associated with it. And it's pretty similar to DA. Like every transaction today pays a little bit of money for posting DA to Ethereum.

In the future, every transaction will pay a little bit of money for DA and also for proving. The proven costs do scale with the number of transactions in the chain, but it's kind of, people should think about it as every transaction paying its little bit to the overall proving costs for the entire roll-up. And those proving costs are really cheap. So for example, for an E or C-20 transaction, that's like 20 or 40K gas, it ends up being like hundreds of a cent. It ends up being less than a tenth of a cent in proving costs. So I think that actually might be comparable to the cost they'll pay for DA when blobs are priced on target. And for, you know, more expensive transactions, like a deck swap, it's like a couple tenths of a cent. And then at the top end of transactions that are millions of gas, it's like around one cent. And although all those costs are honestly pretty cheap compared to like what users on base even pay today, I think the average transaction costs on base is like a couple cents, especially when you factor in priority fees. And so the, I think overall the proving costs actually have really nice scaling properties for very high TBS chains.

Umay, this, uh, you like you said, this is a brand new technology that we are just testing out and all technologies iterate and improve. Can you kind of simulate for us the future of the, the cost structure of this like whole tech stack as it kind of matures, gets better and improves? Like how much, how much juice is left to squeeze here? I think there was a lot. So, you know, six months ago, I think, well, a year ago, people didn't even think this was possible. They thought ZKVMs would be too slow. If you just take RustCode and stick it in a ZKBM and try to prove the EVM, it would be like way, way. It would be impossible.

It wouldn't even finish in time. It would take, like, days. Then a year ago, I think it was like millions of dollars per year to even prove like Ethereum blocks or prove any of this stuff. Then when we first released Syscinct, I think all. all this stuff would have taken like dollars. So it was like 10x more expensive. And since we first released succinct in February and SP1 in February to when we, to today, we've improved our performance and costs by 10x.

So that's pretty significant. And I think over the next year, there will be another 10x that comes out from like better cryptography, better algorithms, you know, even better Rust Code. So overall, I still think there's like a 10x left, if not much more. and hopefully at that point the proving costs are like almost zero so users don't notice. So you're talking about the I've heard this term and that I don't fully understand what it means is like the cost of proving needs to go down before we can do XYZ, right? I hear that repeated about like distant Ethereum futures and all sorts of discussions. Is that effectively what you're talking about, the cost of proving like kind of like going down?

And why does that matter? Why is that significant in the scheme of like unlocking the potential here? And like when you say it goes down, how does it go down? Is this like a Moore's law thing? Are we like improving some hardware somewhere? Like this is something I've not quite understood, but like you guys are the ZK experts. So maybe you could shed some light on this. Yeah, ZK definitely has its own Moore's law where the costs have been going down by 10x like every year.

And that comes from a bunch of different things. It certainly comes in part from like better hardware. Like when we first started, we were using CPUs. Now we use GPUs. It also comes from better cryptography and algorithms. So we're like innovating on the ZK front to make our algorithms just more efficient. And that just comes from like being smarter and like thinking harder about how to make all this stuff theoretically better. And so the combination of those two things is what's leading to these like 10 Xs every year.

And do we expect that we'll continue? It's just kind of like, you know, you see those charts like Moore's Law, of course, is a famous one in compute technology. but there's also like genomic sequencing and sort of the cost to sequence a genome. It's cost billions. And now it's like, I don't know, we're down into the thousands and that will get into hundreds. How much more can we squeeze out of this proven costs like through hardware acceleration and software algo improvements? Do we have, you know, many years of this to expect into the future?

Yeah, I mean, it's happened for like the past five years. So I think there's no reason to think it won't continue. So I'm super optimistic that we still have like orders and magnitude to squeeze out. And I think definitely, again, if you look at SP1 from earlier this year to now, there was like a 10x. I think if you look hopefully by this time next year, there'll be another 10x through a variety of factors. I think one thing to point to is like if you got, if you look at AI, if you look at GP3 versus GPD4 versus like the new GP, whenever they put out GPD5, you kind of see a similar thing where the models just get better and better and better and better. And it's like really astonishing and remarkable. And I think a big reason for that is like those models

had the right form factor where anyone can talk to them in English and they like spit back out, you know, smart words basically. And I think ZKs finally arrived at like the right form factor where you just put in normal code and you get out proofs. And now that we have this like final form factor that's really flexible, that's general purpose that can prove any computation. And you can prove any computation that can prove any software, we can just take that form factor and basically like turbo accelerate it and make it faster and cheaper because it's like it's the correct like final form. So I think that's like the unlock to enable ZK to like be everywhere. Okay. And then when Prover costs get really cheap, as they will, and we're plotting kind of the trajectory of this, then so like what does the

average crypto user get out of that? Do their ZK roll up transactions like get? Get? get cheaper, like 10x per year? Do they become, like, is that what we get? What do what is every, how does everyone benefit in crypto from really cheap proving costs? Yeah, their transactions will get cheaper. They'll have more security. In some cases, they'll upgrade from having zero security and relying on a multi-sac to having actual security. And everything will be verifiable. I think like verifiability is this like key cornerstone of crypto, right? Like the whole point of all the systems we're building is that it's decentralized. is transparent, it's verifiable, and with cheap Z-K, now we can just make a lot more things verifiable.

So every world will be verifiable, every transaction is verifiable. It can all talk to each other. There'll be seamless inoperability. New projects are coming online to the Mantle Layer 2 every single week. Why is this happening? Maybe it's because Mantle has been on the frontier of Layer 2 design architecture since it first started building Mantle DA, powered by technology from EigenDA. Maybe it's because users are coming onto the Mantle Layer 2 to capture some of the highest yields available in Defi.

and to automatically receive the points and tokens being accrued by the $3 billion Mantle treasury in the Mantle reward station. Maybe it's because the Mantle team is one of the most helpful teams to build with, giving you grants, liquidity support, and venture partners to help bootstrap your Mantle application. Maybe it's all of these reasons all put together. So if you're a dev and you want to build on one of the best foundations in crypto, or your user looking to claim some ownership on Mantle's defy apps, click the link in the show notes to getting started with Mantle.

Radically simple ideas always tend to catch on. That's why Cartese did the hard work of putting Linux on chain, so that building DAPs can be radically simple by using Python or JavaScript and their suite of libraries. Simple, like not rebuilding the basics from scratch. Simple, like dedicated scalable compute for your DAP. Simple, like building DAPs however you want. Web 3 should be simple too, like bread and butter. Cartese brings radically simple solutions to Ethereum, so developers can do what they do best.

Build. Go ahead and discover a flexible, modular stack on Cartesee and build your most powerful and ambitious project yet. Visit cartesi.io slash simple and simplify your blockchain journey and start building today. So, Umah, I think, represents the vertical scaling strategy in this conversation. Andrew is the horizontal here. So Andrew Uma just gave us the roadmap for how she's going to allow a chain to scale to its logical conclusion with being verifiable and secure. You are going to distribute this technology. You are the distributor. Can you talk about conduit's role of like democratizing access to all the gigabrain stuff that Uma's building.

For sure. And then one thing that I think wasn't mentioned, but I think is important in kind of the vertical scaling topic is latency. And so right now it takes about 45 minutes. Uma can correct me if I'm off there, but about 45 minutes to generate a proof. When that goes down to like 30 minutes, 15 minutes, even something as real time as like one minute, I think that's where cross-chain communication starts to get really interesting. And today we're relying on basically interop protocols to front capital. when you have the proof that something happened, and you can actually have double proofs on both sides,

I think that starts to get really interesting in terms of kind of a unique primitive and removing trust from the equation. And ultimately, like, what we're building here is, you know, scalable systems that don't require trust and that have verifiability. And I think it's a very important cornerstone,

in addition to, obviously, the cost is the latency of generating kind of these proofs. In terms of distribution, the way that I kind of think about ZK and kind of Ethereum in general, you know, we have so many different role of frameworks, right? And they all have like different ways of communicating their state to Ethereum. And you even probably felt this at the DevCon Interop Summit, right?

Is you have these different frameworks, how do they all kind of talk to each other? And kind of the interface is really, can I withdraw to Ethereum? And with a seven-day kind of optimistic challenge window, it's very difficult and very slow and very expensive to do that. And there are like some solutions, like Super Chain is maybe a good example where they're trying to find a way to get faster kind of native interop, working with an optimistic kind of faultproof. I think one of the, and like there's a good story from AWS where when they were scaling the company, instead of kind of teams reaching into each other's code bases and like making

modifications, they define very clear API boundaries, and that was how AWS kind of scaled. I think ZK with fast finality has the potential to do this for all the chains across Ethereum, right? So you no longer have to introspect the implementation of a particular roll-up. You don't need to know about another rule of state. All you need to know is that you can have fast finality and move capital out of that roll-up, you know, starting with an hour today, maybe it goes to, you know, five to ten minutes in the future. And that being, kind of the ultimate interface where you don't have to know any implementation details, but you can easily move and transact across chains seamlessly kind of on Ethereum.

And so that, I think, more than anything, is kind of what we're excited about is, yeah, again, like derisking and kind of decoupling these chains and enabling a flourishing of innovation and different types of kind of roll-ups across Ethereum. Andrew, I really like that metaphor. I remember hearing it like once before, and I think it's worth kind of going over it again and its significance. The metaphor of like internal AWS teams, this was about like cross-team cooperation, communication,

compatibility with like Amazon working itself as an org structure. And Bezos, the leader here is like learning that like different components of like the AWS team would like different teams would talk to different teams, like teams of five persons talking to another team of like five persons in like a different part of the org. And I think there's another, like, famous Bezos quote of just, like, your team needs to be able to share two pizzas. And if that, if you can't feed your team with two pizzas, then you need to, like,

break apart the team because they're going to, like, slow down. So he liked these, like, small, modular kind of teams. And the story here is that he wanted to have these teams stop talking to each other because that slows down progress. And so there should be, like, API endpoints. These teams, if you want to talk to another team, you got to talk to their API. Like, you don't go talk to the team.

You go talk to their API. And that's what scaled coordination internally to Amazon. And what allowed, like, Amazon to become, like, the gargantuan that it is. And you're extending this metaphor to roll-ups. It's like, stop heart, stop going into, like, the code base of these roll-up teams. Stop looking at, like, what their internals are like so you can figure out how to couple there. Just, it's a ZK-proof, just, like, hook right into that. And this is like extending this metaphor to just like this is just how efficient cross-chain communication can happen faster, lower latency with like stronger assurances.

What can you add to that metaphor? 100%. This is really, I think, the execution of the role of centric roadmap, right? And Ethereum is kind of like, you know, AWS. And like how do we get these different teams or these different parties that are building on top of Ethereum to collaborate and coordinate in a way that provides a seamless kind of interface. and kind of experience for the end user. And I think, like, just to take a very explicit example, like, you know, the optimism and arbitram code bases are diversion in terms of how they approach roll-ups. And it might be tough to get those two teams to kind of work together.

With ZK, like, kind of doesn't matter, right? Because Ethereum is that, like, unified interface. ZK is the technology that enables that kind of fast finality. And it's kind of really the glue between kind of the communication there. And so I think that when they say ZK is the endgame, I think it's really true. I think what we really underrated is how fast it would kind of get here. And I think, you know, credit to SP1 and great teams working in ZK for making that happen. So ZK proofs are kind of like the restful API that allows like all of these different like roll-ups to, you know, communicate with one another as a standard.

That's right. So I would say basically the, it's really the faster finality and the validity proof that's enabling faster communication. And so optimistic roll-ups by their construction with the challenge game and kind of censorship-resistance requirements will always need a very long kind of challenge period and the brawl period. And so this is kind of the core thing is faster finality. Well, it's interesting because whenever we talk to, you know, like one fixing, like fragmentation solution set that's been brought up in recent months is, you know, Polygons ag layer. And whenever I've, like, we've asked about, okay, what's required for every. to participate in like this idea of shared interoperability between all of the rollups and kind of an ag layer type standard, whether it's a polygon standard or something else, like just something that

uses that tech. Always the answer is like, well, only ZK rollups can participate, right? And I think that's effectively what you're saying. But like this gets a fast path to all the optimistic rollups to become ZK rollups. And then they can all interoperate roll up to roll up without going back to Ethereum mainnet in like kind of an ag layer type construct. Is that sort of what you're saying here, Andrew? Yeah, I think there's, so there's going to be like, so Ag layer, I would say, is like a kind of custom implementation that they're going to work on that is going to enable additional features. What I would say is what ZKK provides is a minimal kind of kind of interface and API so that anybody can communicate with anybody.

And so Ag layer is going to be for, you know, AgLayer specific chains in the same way that like, you know, super chain is for super chain specific chains. what ZK enables is kind of the lowest common denominator without kind of high overhead. One potential problem that I might see, Arbitrum loves their orbits, optimism loves their super chain, and you guys are presenting this like alternative solution. And there's that classic XKCD meme of we have 14 different interoperability standards. You know, Polygon wants their ag layer. Now we have this like 15th. And sure, it's better.

But how do you actually convince the market? How do you convince adoption of this standard when optimism has just been so hyper-focused on their own internal system and so has arbitram? I think with superchain or orbit or Aglar, like ZK is not competitive. It just like helps accelerate all those things. So especially with something like Agler, it really relies on every roll-up being ZK. But I think Aglar can actually accommodate any stack. So Aglar, you can have OP stack chains connected to Aglar. with OP Sissinct, you can have any chain that has a ZK-proof connect to Aglare.

So it's actually super general, and it's uniquely enabled by ZK. And also, oh, by the way, they're building Aglar with SP1-2. Same with Superchain. Like, Superchain is kind of two things. One is this, like, protocol for shared sequencing and interop at, like, real time, basically, to really make all the chains of the super chain feel like one. But then there's, like, settlement of all the Superchain chains and making sure that, like, you know, the validity of kind of having these cross-chain

super-chain transactions are preserved. And that can be done with the ZK proof. So superchain one day can use ZK, and that'll just make all their lives better, basically. So it's actually an accelerant, not competitive. One reason why I've been, like, kind of bearish on interoperability between the layer two base in Ethereum, like one hurdle that Ethereum's layer two is it's going to have to have to get over for me to get bullish on endrop was the fact that it was a lot of social coordination more than it was technological coordination. But, um, I think what you're alluding to is like,

there is no social coordination here is actually we could just go back to the tech and it's actually just like adopting superior tech and what results out of that is just better intropability. Is that, is that correct? Yeah, totally. I mean, tech problems are easier to solve than social problems. And I really like what Ryan said about ZK being like the restful API of rollups. Like each roll-up, if I have a proof for an OP roll-up and I have a proof for an arbitram roll-up, then those two roll-ups can talk to each other. And it kind of doesn't matter if, like, the teams don't like each other or, like, you know, they're not building a protocol.

It's just a proof is a proof. I love that. The classic Bitcoin meme about this is Bitcoin is money for enemies. Like, Bitcoin is trusted between enemies. Like two enemies can transact using Bitcoin. And, you know, optimism, arbitram, not enemies. But, you know, they've had their tensions. And so like maybe the SP1, the ZK proofs is cross-chain communication for, you know, teams that are competitive.

Now I want to broaden this for a second to a topic that's also near and dear to my heart and our hearts and everyone listening probably to Bankless who's in the Ethereum space, which is this concept on the more distant Ethereum future roadmap of native ZK rollups. and since we have some ZK experts here, maybe, one of the things in the future and one of the hopes for Ethereum is that the execution layer on the L1 actually develops to the point that we can like snarkify the EVM as like Justin Drake likes to say in addition to like snarkifying the consensus layer. And I'm wondering from your perspective, Uma, what is necessary from a ZK perspective to like allow that to. happen. So can we have a native ZK EVM on the layer one using this type of tech? Would SP1 be involved in that? How close or how far away are we on that? And like another thing that I've heard, which is like, you know, proving costs have to go down in order for this to be actually practical. Anyway, what are the constraints, the bottlenecks? How soon are we to that vision of a ZK roll up on the layer one? Yeah, I think there's a few components. One component is like the proving costs, which actually are pretty

cheap. So I think we did some numbers and we think to prove all of Ethereum for an entire year would be a couple hundred thousand dollars. Like that, well under a million dollars, it would be, that's actually not that cheap at all compared to like the rest of the costs of what people pay on chain. I think the main blocker today for an enshrined CKEVM is latency. If you wanted to be part of the protocol, then you need real-time-proving of Ethereum blocks so that when the proposer proposes a block, they can also include a proof of what the blocks outcome is and a proof of what the next state is. And for that, latency is super important, and we need to be able to generate a proof in less

than 12 seconds of an Ethereum block. So I think the next frontier of ZK is actually not only cost, it's actually like this real-time-proofing latency thing. And I think we're not quite there yet, but I'm pretty optimistic that over the next, you know, year or so we'll be able to hit that in ZK World. You think over the next year or so we'll be able to hit the latency of like, do you say 12 seconds or under to like make this possible? Because that's now the constraints. It's not proving costs, actually. It's just like latency of proving. Yeah.

What else needs to happen for like, is that the only constraint? So we get those two things. I'd imagine some of this has to be more battle tested, because when we're talking about, like, deploying this across like a, you know, $400 billion network stakes are a lot higher. So is that another? Like, how do we know when it's ready? This tech is actually ready to be deployed to main. I'm sure we can experiment in these layer twos, but will that take some time as well? Yeah, I think that's a really good point. The security of all this stuff is super, super important, especially for a really big chain like Ethereum. And I think SP1 helps with that because when you try to prove the execution of an Ethereum block with SP1, you're taking REF and REVM, which is like normal Rust Code and proving it within SP1. So you're not hand-coding this complicated contraption for the EVM. You're just taking normal rust code and proving the EVM with that. So that makes it a lot more secure because like REF is an actual node that a lot of people are running. So it kind of like reuses, it's actually used in practice.

But I think SP1 itself, obviously, you know, that has to be more battle tested. It has to go through audits. Or it already has gone through audits, to be clear, but it has to go through more audits. And then the Ethereum Foundation recently launched this $20 million grant program for formal verification of ZKVMs. And that can also really help with security and making people feel good about stuff. So, yeah, I think you're right that for this to be enshrined in the mainnet layer, there will definitely need to be a lot of conversations and work. around that. Andrew, do you have any takes on this, like native roll-ups? Yeah, I mean, native roll-ups,

I'm less educated on. I think, like, what's nice about what we have today is you have a lot of different opinions and you have a lot of different, like, implementations, and people get to opt into the implementations and the trade-offs that they care about. So I don't know quite as much about, like, enshrining one roll-up implementation. That, to me, seems like a premature optimization. But I think in terms of a native ZK-EVM for layer one, I think that, I think that, is obviously kind of the endgame. I don't quite know when we're going to have that, but we have a lot of discussions today about performance on later one and what it takes to host a very decentralized kind of network. How do we enable the solo staker? And like if we want kind of

our cake and to eat it too, which is having a decentralized network, making it performant, faster blocks, which, you know, is also going to help the kind of roll up centric roadmap. we're going to need to get to ZK because ZK enables you to run a much lighter kind of validator note, right? That is really just proving or just kind of verifying the proof of the execution versus needing to do the entire execution themselves.

And that's today, like ultimately the bottleneck on getting, you know, more folks kind of validating the network. So to me, ZK is kind of the endgame in almost every capacity, I think. The core question is like, when does it get cheap enough? When does the latency get fast enough to really enable the types of kind of real-time proving that we need to enable that. As we wrap this up and thank you for your time with Uma and Andrew, this has been fantastic.

As we maybe bring this to a close, I want to ask you guys a broader question because, like, so when we do episodes like this, I get even more excited about Ethereum and the roll-up-centric Red Mac because the end game of ZK deployed across roll-ups now sounds like it's like much faster than I anticipated. And that's because the ecosystem is working. parallel across like many of these things and like uma like what looks to us like an overnight success and suddenly it's here and we can just convert our optimistic rollups into zk rollups. I know you've been working like tirelessly for years on this.

So it's like probably doesn't seem as surprising. But to me it's surprising. I want to ask you guys a question about the sentiment and narrative around Ethereum because I'm not going to lie. It's not great right now. Okay. Price is down. People aren't feeling good.

You know, David mentioned parasitic. two's and you guys have heard that narrative. But in general, people are like fading Ethereum and the vision here and like the roll up-up-centric roadmap and the choices and the product decisions that it's made. What's your take on this? Maybe I'll just open it up in general to your take on the really bad sentiment around Ethereum right now. First, you, Uma. Well, I think the road-centric roadmap is like totally correct. I think if we want to build the world computer, it's not going to fit all one machine. I think that's totally ridiculous. And that is the Solonner approach. And I think it has some tradeoffs in the short term, where it like short term

seems good, it's working. But in the long term, if we really want to put all compute on chain, it's very obvious that it has to be distributed across many machines. And the only way to do that in a verifiable way is with ZK. So I think to me, actually, the roll up center roadmap is exactly correct. It's like to maximally scale. If you look at any other system in any computing, like, paradigm, you've had this, like, horizontal scaling. And the way we do horizontal scaling in blockchain, that's verifiable, is with ZK. So I'm actually, I think the core thesis is correct. I feel like though, right now, there's a lot of, like, growing pains. We're going through, like, our adolescence phase, where the roll-ups are actually, now they're working and, like, now they have users. And now there's

many different teams that are battling it out. And they're not talking. to each other and it's kind of awkward. So I think like actually having fighting in the roll up-up-centric roadmap is a sign that there's actually something to fight over, which is, I think, actually good. I do think we do have to get to the point, though, where we kind of get out of that phase and get to everything being really high TPS and everyone talking to each other and not having this like kind of awkward growing pains again. So that's like the thing I'm excited about over like the next year is like actually getting

to unification and. interoperability across the roll-ups, and I think ZK will be a key part of that. But I think the fundamental roadmap thesis is actually totally correct and will actually scale and stand the test of time. I agree with a lot of that. I think that, you know, we're still so early in kind of the roll-up centric roadmap, right? We launched Blobs like six months ago. And, you know, they launched and they're free and we're starting to finally see kind of price discovery there. To me, it seems a little myopic to look at the fees generated off of like a new product and to not think about like three to four or five years from now.

And in some ways, like, I always like to show people rollup at WTF in terms of the amount of TPS and like MGAS a second we're doing. You guys put that together, right? It's great. We're just looking at that with recording with Justin Drake. Yeah, it's a great performance-based dashboard. But I think what it shows is one, we've scaled a three in like 50X or so. But if you compare like 300, 400 TPS with like Visa alone, right, does like 50,000 TPS.

And if you look at Facebook, Instagram, they're doing like millions or, how. hundreds of millions of TPS, just the amount of stuff we're doing is tiny. And so to think that this is the end state and that this is all that there is, to me, is very myopic on crypto in general. And like we believe in a world in which, you know, on-chain compute is really abundant and that the role of centric roadmap is the way to achieve that in kind of a secure decentralized way. And so ultimately, you know, there's going to be a tradeoff of spectrums. Ethereum has always been very kind of, you know, on the censorship resistance and kind of decentralization side of things. And, you know, we believe that if you're building crypto, you know, those are

some pretty important primitives. And it takes time to scale it out. But I think with innovations like SP1, we're getting there. Still bullish on the world computer. I'm bullish. This is great. Might take some time, though, I guess. Umah, Andrew, thank you so much for talking to us all about ZK and exciting developments. Thanks for having us. Thanks for having us. Bankless Nation, got to let you know, crypto is risky. You could lose what you put in, but we are headed west. This is the frontier. It's not for everyone, but we're glad you're with us on the bankless journey. Thanks a lot.