Taiko: Scaling Ethereum in a Decentralised Manner - Joaquin Mendes. Ep. 597

Transcript

If you ask me, I think we are at 1% of the potential usage that I expect Ethereum and its rollups will happen in the future. If we expect Ethereum to be the new internet and I know this is like a buzz word, but if we expect that to happen and if we expect that the blockchain is is completely abstracted like the internet is abstracted today. uh I think we need a lot of scalability on the layer one in order to you know attend the demand that happens on the layer one but also to attend all the demand that comes from the layer two base sequencing means that the sequencing is not done by by us by tao we run one sequencer but it's permissionless as anyone can do it it's done by layer 1 validators welcome to epicenter the show which talks about the

technologies projects and people driving decentralization and the blockchain revolution I am Fedria ans and today I'm speaking with Drain Mendes, the COO at Tao, which is a based um layer 2 on Ethereum. Before I talk with Drain, um let me quickly tell you about our sponsors this week. If you're looking to stake your crypto with confidence, look no further than Chorus 1. More than 150,000 delegators, including institutions like BitGo, Panta Capital, and Ledger, trust Chorus One with their assets. They support over 50 blockchains and are leaders in governance on networks like Cosmos, ensuring your stake is responsibly managed. Thanks to their advanced me research, you can also enjoy the highest staking rewards. You

can stake directly from your preferred wallet, set up a white label node, restake your assets on layer or symbiotic or use the SDK for multi-chain staking in your app. Learn more at chorus.1 and start staking today. This episode is proudly brought to you by Nosis, a collective dedicated to advancing a decentralized future. Nosis leads innovation with circles, Nosis Pay, and Metri, reshaping open banking and money with Hashi and Nosis VPN. They're building a more resilient privacy focused internet. If you're looking for an L1 to launch your project, Nosis Chain offers the same development environment as Ethereum with lower transaction fees. It's supported

by over 200,000 validators, making Nosis chain a reliable and credibly neutral foundation for your applications. Nosis DAO drives Nosis governance where every voice matters. Join the Nosis community in the Nosis DAO forum today. Deploy on the EVM compatible Nosis chain or secure the network with just one GNO and affordable hardware. Start your decentralization journey today at nosis.io. IO. Hey Jin, thank you for coming on. Hey, thanks for having me.

So, I've been following Tao for quite some time. Um, but before we kind of dive into kind of like the nitty-gritty technical details, tell me about yourself and how you came to be um in the web 3 space. Okay. All right. Um, so as you mentioned, my name is Hakin. I'm COO at Tao. I've been working uh for Tao for the last couple years. Um before I used to be head of partnerships. So on the early test net days uh I started to bootstrap the ecosystem and now recently like half a year ago I I jump into the co position uh for which I essentially drive the business strategy also more boring stuff such as pure operations but

uh but mostly mostly growth and and strategic growth. Um my journey in crypto started I would say five five or six years ago. Um not professionally but more like as a hobby I guess like like like many people I started just trading and and listening to podcast and and understanding the technology. Uh but professionally it started um around three and a half or four years ago. I I started working at Polygon Labs um doing mainly business development uh for a year and a half. I did some chief of staff for one of the co-founders but that was uh that was a short period

though it it let me see some you know some good things on on how these kind of startups are driven. I came from corporate world from PWC which is like a big big corporate world uh doing consultancy. So actually you know this helped me understand how how a crypto startup runs and it helped me on my on my current role right. So yeah one year and a half or to two years in polygon then a little bit more than two years in Tao that's my journey. Mhm. Tell us about Tao and how Tao got started.

So, Tao got started um I don't know the exact dates actually I should I should know but around three plus years ago uh but it actually we can actually say it started earlier because the founders of Tao are actually the founders of Loopring and Loopring uh for for the audience that doesn't know about it was the first uh CK roll up on Ethereum Um, it wasn't a general purpose one. It was an application specific uh for It started as a DEX, right? Exactly.

Correct. We had them on ages ago. I was super um I I was uh Yeah, I was uh really bullish on on the tech. Uh it it was a very cool application. Exactly. At the time it was it was super noble. Uh it was it was pushing the the boundaries. Uh actually was founded in 2018. So, so yeah, we talked with them a lot because kind of like we we kind of um we we built cow swap and so kind of which at the time was Nosis protocol. So kind of obviously kind of like there was a lot of kind of similarities and kind of like they they ended up focusing on different things than we focused on. But yeah, there was there was definitely um a lot of uh mutual appreciation um and

admiration. Interesting. That's very that's very interesting. Yeah, it was it was wild times. That was that's actually when I started in crypto more or less and uh yeah it was it was very noble at the time right uh scaling Ethereum no one was building this this roll up thing and um so the team at Tao the founders at Tao have a lot of experience on the scaling space. Um what they found out was that uh well then the in 2020 or 2021 um you know some rollups and side chains they started to pop up and scaling Ethereum was um was uh was a normal thing and um they realized that all the

solutions that were coming up to the market for scaling Ethereum uh were not really like super aligned with Ethereum which is normal right the EVM the execution environment is quite complex so try to achieve compatibility or even equivalence with the EVM with the EVM is really hard so some projects you know that that started popping up at a time um they made some com some shortcuts some compromises which is again which is normal but Tao was born with the idea of um of making you know as little comprom compromises as possible to to Ethereum when scaling it. So it was born

initially as a type one CBM. Type one means uh the highest level of compatibility. It's actually called Ethereum equivalence. Um so the idea is that developers don't have to change their code when they migrate from Ethereum to Tao. um which has a lot of benefits of course but has a lot of um you know obstacles when it comes to to building the protocol uh for for obvious reasons I guess and so it was born as a type one CVM now we have evolved the concept as the as the technology have evolved to and now we call ourselves a base rollup the first base rollup on

Ethereum which TLDDR and I am sure we will we will expand on that later means we are using Ethereum layer one for sequencing instead of having a centralized sequency doing all the job of of uh you know ordering the transactions into blocks. Yeah, absolutely. I think that was that that was a very comprehensive introduction. Maybe let's kind of dive into the um type one ZKVM um thing first because that's that's kind of neatly self-contained and then we can go into the sequencing later. Does that work for you? Sure. Sure. So kind of a ZKVM um kind of what it does kind of like and correct me if I'm if I'm mistaken here, but kind of like um it just takes all the op codes that we have on Ethereum

and kind of just transposes them one to one one to kind of like a ZK circuit, right? So Correct. And I I remember kind of like back in the day when we thought this was impossible and then kind of Judy came along and kind of like he he kind of talked about this at ECC and everyone was totally flashed. So maybe tell us a little bit about the provenence of kind of like this idea behind um behind kind of uh transposing all the uh known op codes into ZK um circuits um and what the challenges are and why why we used to think this is not possible.

So yeah, um you defined very well. Um actually when I when I when I was at Polygon, I I had a chance to meet Jordi Jordi Bina and he's actually from Barcelona and from Madrid. So so we we we got along well. Um, one of the co-founders that I used to work with as chief of staff was Antony Martin and he he was actually one of the founders of Hermes, which was the the then the Polygon CKGBM when they when they acquired it, but that's another story. So, um, essentially that's exactly what you mentioned, right? Uh there's um four types of uh of uh CK CK EVMS as um that's simply a taxonomy that Vitalip

made from type one to type four. Actually this is not very much used as of today and I will explain why but uh back in the day there there was there was considered there were four types. There are some some rollups in between two types but that's a different story again. Um type four means um you are essential for example CK sync was a type four um they made some changes to the AVM in order to make it more performant. So they are not equivalent down to the op code level. They made these changes uh for um obvious reason. The EDM was not created in the beginning to to to

support the roll of functions. Let's put it this way. Um, so it's not friendly. It's not friendly for for rollup development. So, uh, the closer you get to the type one, the more sacrifices you make on performance. Okay. But the closer you are to Ethereum, the type one is the only one that is equivalent down to the op code level. Uh, as you mentioned before, so uh it is actually it was actually a trade-off. Um, I'm not saying being type one is better or being type four is better or being in between is better. It's just a trade-off. If you get farther from the type one or closer to the type four, you are optimizing on performance but sacrificing

compatibility. Um, more performance obviously is is better for for some use cases and some some um contexts. Uh, and more compatibility is better for other stuff. You can reuse the tooling. uh you can uh you for some applications you don't have to perform extra security audits when you migrate from Ethereum. So from a security standpoint is kind of like better but then you know performance is also important especially on rollups. So it is it is a trade-off. I don't know if this answers your question. Yeah it it does answer my question. So kind of the rationale behind this was that um compute would be less of a bottleneck than composability.

Is that is that kind of like a fair oneline summary? It's a fair oneline summary because uh I think the founders uh and and and when I learned about Tao what we expected is that um you know all the hardware and the software requirements would would accelerate that much that um compatibility like even being on the on the on the closest end of uh of compatibility is not going to be an issue in the future because scalability will be solved at the layer one and then the the rollups the rollups that will win are the ones that are truly truly truly aligned with Ethereum because it it enables a lot of different stuff that that we can talk about later.

Are there any edge case op codes that were especially hard to replicate in a ZK circuit? Because kind of like I remember times when it wasn't even clear whether you in principle could replicate every single op code in a ZK circuit. So kind of like te tell us about the um the levels of the the varying degrees of difficulty in kind of like transposing these up codes. I'm not aware of a specific examples. what I know about is that um Tao started like as I mentioned like three three plus years ago and um if you think about it like if you replicate um an execution environment that already

exist it seems like easier right than than than than creating something new like for example seeking deep but I know for a fact and and and I'm not a developer but I I speak every day with the developers of of Tao and and and other folks in the space that are on the on the on the more technical side. I know for a fact that um probably there were if I had to mention one example which I I don't know technically uh I I know I would be I would be leaving out a lot of others. I know there were a lot like quite a lot of challenges when it came to to achieving the total equivalence down to the up code level and um and that's why it took a little

bit longer than expected to develop uh to develop the roll up and that's also why uh most of the rollup apps um and they they also say publicly I'm not I'm not putting putting words in anyone's mouth they like to start on on on type on higher types on type three type four sometimes Even type two from the get-go like I think scrolled it uh but they never start in in the type one. They like to progress progressively go to type one because it's because it's it's harder. How do the um Tao ZK circuits compare to the ZK EVM uh that Polygon built? So kind of are they the same circuits? Do do you kind of reuse this um or do you

kind of come up with your own circuits? I believe uh Polygon created a a different circuit than the others. I I don't remember exactly the name. I think it was Cir. Uh but but I'm not entirely sure. I'm I'm sure that we don't we don't use uh we don't use the same like that's for sure actually. Okay, let me rephrase this. We didn't use the same because we are no longer a CKVM. And I think this is important to mention. Yeah, we we'll get there in just a bit. Yeah. Yeah. we transition from CKVM to CKVM um for for many reasons um so we don't use the same circuit as I think we are unique in that in that in

that sense perhaps we are we are we are similar to other type shoes like scroll but certainly not like polygon which uh frog was polygon ckvm of course not the poss which is a side chain they were a type three from the get-go and and and we we were a type So not the same. Can you describe to us what base sequencing is or what's kind of understood by this time? Uh by by this term, sorry. Sure. Um base sequencing means that the let's define the sequencing, right? Sequencing means um ordering the transactions and creating blocks. Okay. Um the power of sequencing is actually pretty big. uh if we think

about the all the me you can extract like it's not trivial how how transactions are ordered into a block of course right um in most of the cases if not all um the sequencing is done um on a centralized manner I'm talking about rollups okay um sequencing is done in a centralized manner by a centralized um party or hardware to put in a way base roll up um base sequencing means that the that the sequencing is not done by by us by Tao. We run one sequencer but it's permissionless as anyone can do it. Um it's done by layer one validators. So not anyone like layer

one validators can can sequence the transactions on the layer two and this has some pros and cons. Uh this unlocks some benefits but also has some drawbacks. Do you think it's fine if I go through those now? Yeah, absolutely. Yeah, I mean maybe for that um tell us how the um is it an optin thing kind of like do all L layer one kind of validators kind of sequence blocks for Tao or is it something that they kind of have to choose to do kind of like you would in reststicking? They have to choose to they have to run the the Tao nodes. Um they and what percentage of Ethereum nodes run the Tao node? Not not not many now. And I I will explain now like very

very few uh actually at our peak we had like u 70 something uh validators proposing blocks on tao but there's an explanation to it because the because and I I may maybe if I if I talk about the drawbacks now I can I can explain why why it happens. Okay. So um I will start with the drawbacks and then I will go through the benefits. Um main drawbacks of uh of uh base sequencing are uh one of them is um on profitability and profitability we can unpack this in in two manners. Um first of all Tao doesn't capture the revenue

from the fees. uh there's other rollups that are making that are making a lot of money which is fine uh out of uh out of sequencing the transactions. We don't capture that that value. So that's kind of like a drawback if you think um of a of a base roll up as a business. Um it's harder for us to make to make a profit in that sense. Um but also um it's uh it's a little bit harder for proposers on Tao to make to make a profit and I will explain why. Uh the reason why um layer 1 validators haven't chime in in Tao uh to to to propose blocks is because it is

hard to to make a profit right now out of proposing blocks on Tao. Um the main reasons why without getting into much detail is um TLDDR we use the Ethereum block space in a less efficient manner compared to other to other rollups which means um it's kind of like it's kind of like more expensive for us like other rollups right now they are batching um a lot of of course they batch a lot of transactions together and and and create a rollup out of them create a batch of transactions that what are That's what a rollup does. But they also batch several blocks together before before they send them to Ethereum. In our case, because

we are layer one sequence, we don't have the power to do that in a centralized manner and that complicates things. Okay. And so far, we have been essentially proposing each block separately to Ethereum. So we are consuming more blob space from Ethereum and that makes it um that makes the cost of proposing the block. Every time you propose a block on Tao you gather the fees from the users on Tao but then you have to pay the layer one and if you cannot batch blocks together it essentially means that you are you are paying more per block. You are paying more per transaction on Tao. And that's and that's a that's a very good

question. So the vlog time on Tao right now is around um it it varies but it's around from 30 seconds more or less to from well to 36 seconds to 1 minute more or less. So you have to kind of send an Ethereum transaction every half minute to minute. Correct. And that's that is a that is the second issue right there is there is a UX problem. Um, it's not that it's not only that the block time is is is so big. It's also that the users have to wait that long.

When they're using Tao today, they have to wait. If you go to to Tao, you perform a simple swap, you probably have to wait 30 plus seconds, which is which is really bad. Um the reason why it's 30 plus seconds um and not even more or not even less um is it's it's not it's not okay so the minimum time we we we can wait is 12 seconds which is the block time of Ethereum because we are layer layer one sequence okay that's the minimum but it actually takes longer the reason why it takes longer is because two things um tao as I mentioned before they we run one of the proposers, right? Um, right

now, uh, with the current setup and with the current costs that we're paying to layer 1, the only way to make a profit from a block is when the block is like completely completely full of transactions and and and and we actually propose some very fat blocks like 1,500 transactions per block or something like that. Right now, the activity is a little bit lower, but we have have proposed that kind of fat blocks. the time it takes to fill a block with the correct activity, we are averaging around 2 million transactions per day. It takes around uh one to two minutes, which is even more than the 30 seconds that that that I mentioned, right? So if if Taico proposer wouldn't chime in um it it would mean that uh we would have

to wait one to two minutes every time the block is prop. If we make it completely permissionless and we let only the layer 1 validators propose the blocks, they will only do it when they're making a profit and they are only making a profit when the block is completely full of transactions. And the block is only completely full of transactions after 1 minute or 2 minutes. So that means the UX is even worse. Users would need to wait much more. What we're doing at Tao is we're sacrificing ourselves, our treasury to put it in a way. And we are like like a bot. we are from 30 plus seconds to 1 minute we just push blocks all the time even if they are empty so we some days we are losing money just to ensure that

there is a decent UX so that's the main reason why layer 1 validators haven't chime in yet because um only a few days not not not only a few days but I would say less than half of the days of Tao's existence ta tao has been on mainet for almost one year less and half of the days it was um naturally profitable when Ethereum was super cheap on the layer one and when we had a lot a lot a lot of transactions coming then they they chime in and propose but if not it's it's hard for them to make a profit. Um, have you thought about kind of like raising fees on Tao? Because that's kind of like the the the other lever that you can move, right? We have experimented with that, but we can only raise fees on our

proposer. So for the blocks that are proposed by Tao, uh the the the users will have to pay the fee that we that we put or the transaction would be sitting in the mele scenes forever. But there's always a chance that a community proposers chime in. they can do whatever is fully permissionless and they can they can they can pick up those transactions that that come at a cheaper priority fee and and get proposed. So it's it's very tricky. It's it's tricky. But by the way, this this this will be solved. It's got the same fee model as Ethereum, right? Kind of you have the base fee and then you have the the the priority fee. EIP EIP 1559, correct?

Base fee base fee gets burned. Uh in this case, it's not technically burned. it gets sent to the tao treasury. That's one of the few ways that we capture value from from the fees. Base fee is like um almost nothing compared to the priority fee. Everything comes in the priority fee. A priority fee is what goes to the to the proposer aka layer one validator. All these problems that I mentioned before will be will be solved very soon. But we will get to that. Yeah. Maybe kind of like to complete this. So um you you guys kind of um you commit every block to Ethereum L1 and you actually use the blob space for data availability, right? Kind of like te tell us about I mean obviously kind of

these kind of um it would be weird to kind of commit every block uh uh commit every block to L1 directly and not use the blob space. But tell us about kind of the cost implications um of that and kind of whether you have um uh whether you have kind of an escalation mechanism if that gets too expensive, right? So kind of because for um uh for um the the the transaction that kind of you're um sending to L1 kind of that is a um that's a onetime fee or one time per block fee that kind of you have to pay and obviously depends on what the gas cost is um on Ethereum at that point but

kind of the kind of committing all the um all the transactions to blobs that must be very costly. Correct. It is. So to to to go part by part um indeed we use um Ethereum DA we actually use all the all almost all the services that Ethereum offers possible right if if we if we remember that the taxonomy right you use the settlement you become a layer 2 you use data availability you become a rollup if you use an ali alt data availability layer you're not a rollup actually you're a valid right we also use the sequencing that's why we are a base rollup we don't use the execution ution yet which would make us

a a native roll up but because that's not possible with the current technology but um to answer your question yes we use we use the Ethereum D8 um not only for the for the obvious that that that I already mentioned but also because um if we didn't use Ethereum DA we could be breaking the composability that we are trying to achieve with Ethereum and we will get to that um we have found some bottlenecks actually um when Tao when Tao launched in in in May last year, end of May um we started with low activity but then we from the ecosystem side that I was leading at the time we did a lot of

activations and and we started to have like a lot of transactions actually we picked 5 million transactions in one day um which is a lot but considering how we use the blob space on Ethereum that also meant that we were only Tao via by by ourselves because we were using like half of the of the blob space and then the other half was distributed across the the all the rest of the rollups and then we were pushing the limit of the of the blob space and we were making we were making it it more expensive right well we and all the rest of the rollups of course um in that time what we did um we offer the possibility of use either blobs or call data um because sometimes

cold data was was even cheaper um we don't think it's going to be the case anymore especially with the upgrades that are coming to Ethereum starting with Spectra. We're going to duplicate the blow space and this is just the beginning. So yeah, we expect we we plan to continue using uh Ethereum DA. We have faced uh situations where the DDA was was was way way way expensive. uh we still um used it and um yeah we have the good news coming from Ethereum Foundation that they they are planning to scale the layer one which is going to be super beneficial for us and for our future. Does that answer your question?

Kind of but I kind of I want to I want to push back there a little bit. So you you pass on all the costs to the user, right? So kind of you don't subsidize the blob space, do you? Um so okay not really uh on on our proposer we didn't raise the fees uh to to offset that what happened is we we spent more we we we we wasted more money. Yeah. Okay. Yeah. Yeah. Kind of like when you zoom out what do you think Ethereum um is going to be used for in say two three four five years? Um, and what kind of level of throughput um, do you think Ethereum needs for that? Um,

and kind of if kind of if you kind of post um, all transactions on L2's two blobs, do you think kind of that's sustainable? Okay. Um, I think the layer one itself uh, there's a lot of debate around this. There's no right answer to this question, right? My opinion, my opinion is that um there's different ways of seeing this, right? Like right now, unfortunately, 2025 is still the blockchain is used by um 95% doing DeFi, doing far farming airdrops um you know, it's not

mainstream right? Um so the current scalability that Ethereum that rollups can offer seems like enough right like not many not not many times we have we have pushed Ethereum to the limits but like the if you ask me I think we are at 1% of the potential um usage that I expect Ethereum and its rollups will happen in the future. If we expect Ethereum to be the the the the the new internet and I know this is like a buzz word but but I really don't know how to say it in a different man. truly believe Ethereum can be the home for for the internet moving forward in in many many different

use cases, right? Um if we expect that to happen and if we expect that the blockchain is is completely abstracted like the internet is abstracted today. uh I think we need a lot of scalability on the layer one in order not only in order to you know attend the the the demand that happens on the layer one but also to attend all the demand that comes from the layer two which which I think is going to be what uh what um occupies most of the block space on the layer one. If you ask me about potential use cases that will happen on Ethereum and its rollups, um I think my answer can be a little bit uh disappointing, but this is what I

truly believe. I truly believe uh that tokenization is going to be uh a real thing. uh and but tokenization is so broad but that's how I envision the the the the internet value the money on the internet and and and everything will will happen on a blockchain via tokenized assets and not only for you know tokenized sounds like speculation all the time but tokenized can be can be much more than that right can can be purely efficiency um like compared to the to the current systems like easy example uh uh private bond. Uh how how how you raise funds on a private manner issuing a bond. Um if you compare the web two experience versus web 3, there's

it's a no-brainer. It's a no-brainer. There's no way it doesn't happen in my opinion in the future. But also, for example, I think um Ethereum will be Ethereum or blockchains, however you want to put it, a big home for games. And games is is at is at its lowest point, I think, right now. web free gaming in in in their reputation. But I like I don't I like to think it from a utility perspective more than a real data perspective. Even though most of the transactions today in blockchain happen on games, I think most of them are farming, pure farming, that's not what the gaming wants, right? You want to have fun and then if you can earn something, oh, cool. That's that's on

top. But you don't play to to earn. I don't think that that that that is sustainable in the long term. But if I think about the future, if we completely abstract the blockchain and blah blah blah all these all these all these buzzwords, if a user can have the same game in web two and in web three and and and but in web three they can have ownership and the user experience is exactly the same, I don't think anyone will choose to go uh the web two route, right? So that's also why I think gaming is is one of the things that will for sure and gaming industry is is is massive right if we on board all the gamers we need we need more scalability on layer one like right away for sure and also on the layer twos um I don't

disagree so I think kind of like um blockchains are 100% slated to kind of um be the substrate for the internet of value and kind of like I don't just mean kind of like financial trans transactions but also kind of things that kind have value that kind of we don't technically think of um as financial transactions. So kind of like things to do with identity and reputation and kind of games and so on. Uh 100%. um kind of where I'm pushing back is um if you look at the Ethereum road map um do you see enough throughput there to kind of cover this in

any kind of like even if you say kind of like okay and we we won't we won't take kind of like um we won't settle all the transactions on Ethereum but kind of like we still kind of want this to be kind of like a significant part of the glob with settlement. Do do you think kind of like the Ethereum road map kind of is amenable to that? Exactly. So So what I was getting is that if you want to have all this value from not only financial but also uh like um all the industries that can benefit from blockchain, you definitely need a a big scalability. So if you ask me about the road map, yes, I see I see that Ethereum has a road map

that can fit all these all these um all this demand. Okay. Um are they going to be able to to to make it a reality from a technical standpoint? I don't know. I'm not a developer. I'm not I work for the Ethereum Foundation. But if if we hear the latest right uh not only from Pectra Fusaka and the recent announcement to potentially transition from the EBM to risk 5 it seems what the the numbers that they are that the you know the Justin DS of the Ethereum Foundation are talking about around 10k TPS on the layer one um I don't even remember what they said on layer 2 a million TPS something like that I don't

know but it's in the technical road map to have enough uh throughput to handle all these operations. Will it happen or not? I can tell. I don't know. Okay. Um maybe kind of like while we are kind of I I want to come back to Tao in just a bit, but maybe kind of like while we on the bigger picture, let's talk about um the L2 space as a whole. So kind of kind of you you we were talking about kind of like the Ethereum space and kind of we meant we included all the various L2s on Ethereum in that but kind of like we face a really difficult problem set here as well in that kind of while we kind of

talk about this as one big ecosystem it really isn't. It's kind of it's very fragmented and liquidity is fragmented between them. Kind of like if you if you if you um uh want to exit an an asset from say um optimism and you want to take it to arbitrum and then back kind of like it'll take you it'll take you two weeks return right kind of like could you could you could kind of like mail a USB stick um in the same time. So kind of the that um fragmentation and kind of this um this lack of interoperability between L2s on Ethereum. And I mean that was very much kind of like the idea behind why we wanted Elus um in the first place

because in principle um kind of you could have communication between the elus that was uh you know at block times um and trustless um and kind of we don't have that. What's the problem kind of like in your definition and how do we tackle this? Do we tackle this? That's that's a very good question. Uh first of all, I think um if you if you look at a big picture, right, and how Ethereum has been evolving, I think uh it is it is a beautiful path in the sense that okay, we have one problem. We fix it in a way, but that fix generates a new problem that is that is perhaps less important

than the initial one. then this fix will generate another one and and step by step we keep fixing until we get closer to the end game. I want to say that um I want to give before jumping into the problems that this these layer shoes are are are generating. I want to actually give a shout out to to them, right? Because um every single from optimistic rollup to seek rollups um and I'm I'm very critical with the current setup of of rollups and with the optimisms and arbitrums um on on on X.

I'm I'm critical for a reason, but I also appreciate a lot the value that they have brought to the to the to the you know to the industry to Ethereum's ecosystem. Ethereum a while ago they had we had a big problem of scalability. We all remember the cryptokitties uh thing and and you know how you needed to pay 100 bucks for a transaction in Ethereum. So these protocols came in with the rollups um they did it the best way they knew and they solved that scalability issue and now boom we have uh we can process transactions on Ethereum. It's not on Ethereum but more or less uh on top of Ethereum uh for sense. Yes, that's that's that's a good solution.

But now the problem is that what you mentioned right we have a bunch of rollups um that don't communicate to each other. And it's not only that they communicate don't communate communicate to each other and that you have to wait a couple weeks in the case of of optimistic rollups to to come back and forth. It's also that the liquidity is completely fragmented across all of them and that's a big problem right because uh you know uh if you go to Solana you say okay where do I build I build in Solana full stop that's it easy to understand easy to understand for the developer and easy to understand for the

user and easy to understand for the trader uh that knows where the value is captured easy to understand for everyone it's an easy narrative if you go to Ethereum, you have a big question. Where do I build? Do I build on arbitrum? Do I build on optimism? Do I build on any of the optimism superchain cluster? Do I build on polio? Who captures the value? Um why uh Ethereum layer one is not capturing much value from this or from the other? It is a problem, right? But getting to the point and to answer your question, I think we have solved the scalability issue. Now we are facing the new problem which is which is the fragmentation.

Right now all the rollups are trying to fix it within their own clusters. So instead of having a bunch of smaller islands that don't communicate with each other, they are trying to merge some of those islands into bigger ones but still they are islands, right? So we have the the arbitrum I forgot the name the orbit orbit chase. I think I think that's the name. Optimism super chains sync elastic chains they're trying to make interoperability within those clusters which is very good but still they don't communicate with with each other and most importantly and this is the main reason why I think this is not why I'm critic and I think this is not the

long-term solution is that they don't they are not composable with the most important chain which is Ethereum they are only interoperable within each other but they are not interoperable or composable with Ethereum. The only way to be well this is another password that Justin Drake has been saying a lot universal synchronous composability. Okay, why is that? Synchronous composability is like the highest degree of interoperability that you can achieve. The only way to have universal synchronous composibility, you need two elements.

One of them is having a SE sequencer and the other one is having realtime CK proofs. Okay, if we go to the first one S sequencer, right, you can achieve interoperability within clusters having the the same S sequencer. But if you are not a base rollup, you will never share the sequencer with Ethereum, which means you will never share the liquidity with Ethereum. You will never be composable with Ethereum. You will never have the highest level of interoperability with Ethereum. That's a big problem long term. I don't envision the Ethereum ecosystem longterm without synchronous composibility with Ethereum. That's one thing. Second thing, instant instant or realtime CT proofs. This is the other

condition to achieve the synchronous composibility. This is not possible with today's technology. That's why universal synchronous composibility doesn't exist today. So that's why these solutions that these rollups are proposing are very good for the short term because they are kind of fixing with other um standards the interoperability and optimism recently uh announced how they are making all the super chains interoperable. Please don't ask me about it because I don't know how they're doing it. Uh but I know they announced it. But I consider all these solutions short-term solutions. The only solution long term that I can think of that can

actually make u Ethereum feel like a whole solve the problem of a um you know uh I'm I'm I'm I'm transacting on Ethereum and I I don't even know if I'm doing it on on Tao on Ethereum or on another base rollup that launch as part of the horizontal scaling that we are that we are proposing. The only way to make it feel a one single chain again is through base rollup plus native rollups but that's that's a different story. But base rollups are inherently more expensive than kind of like some of the others, right? Kind of like because kind of you use all the Ethereum infrastructure and that kind of comes at a cost because kind of like it's scarce and even if you make it scale, it's still scarcer than kind of the

equivalent. say kind of like you use um kind of you use Ien DA or kind of you use uh the ABC uh rollup stack from Gelato and so on kind of like you'll always be cheaper than kind of using everything Ethereum natively, right? So what what what's your um what's your answer to kind of which kind of applications will have that highest level of security and for for which applications is it fine to kind of compromise on the trustlessness and permissionlessness um that kind of that you don't you don't necessarily need but then you also don't want to pay for

right so kind of like in my in my head this kind of has to kind of play out as a spectrum where kind of I as a DAP developer kind of have to decide what's kind of what what's um what's the security standard I need and what's the price point I'm willing to pay for that. Yeah, that's a good point. I think there's some applications that are definitely willing to compromise on that. Uh for example, games, right? You don't need the maximum level of security u censorship resistance uh decentralization for in-game transactions, right? You may, but you may want to have it for um DeFi transactions or or or for institutional use cases or for tokenization use cases.

You don't want to you don't you don't want to face the risk of being censored or or or or for the chain to halt. You know, when you have one sequencer that does everything, you you you have a single point of failure, right? There are some use cases that you know where the the lower the transaction value, the less you care about about the chain you you you you launch it on. But to answer your your your first comment on the on on base rollups are more expensive. Uh by definition it seems like they are. I want to jump into the pre-confirmations topic because it is unclear once once base preconfirmations arrive but it seems like uh they can be a little bit more expensive. But look, even if the in

the darkest days where we were um you know using more blob space than the Ethereum limit like consistently I think users were paying less than 10 cents on Tao um which is still cheap and this is just the beginning right if you if we look that's why I say that all these solutions that offer like subsent transactions and and subsecond confirmation times are are short-term solutions but if we think in the long term Ethereum itself is going to it's planning to scale massively not only on the gas limits but also on the blob space and I think the price is not is not even going to be a conversation it's going to be like a commodity that no one

no one will care about so so yeah I don't think that limit is going to be is going to be a problem of the base rollups I think the the benefits especially the composability um is going to is going to be a a game changer a decision maker for both users and developers. Cool. Maybe let's loop back to kind of um explaining the other half of the Tao tech stack. Right. So kind of we've talked about um the type one ZKVM uh pretty uh pretty exhaustively. So let's talk about um the security model of the sequencer and kind of like how you kind of sequence the blocks and kind of how

you commit them um to L1 because you have a very interesting hybrid model here. So on the sequencing side, from a technical standpoint, I'm I'm I'm not aware of anything like super special of of how how the ordering happens in terms of security. I think uh the the the beauty of or or what brings security on the sequencing side is that it is um is that it is permissionless. I think being permissionless is makes by definition not by definition but can be can make by definition something more secure if it if it also comes with decentralization

and decentralization also means uh um you know eliminating single points of failure right that's what makes something something more secure but I would say um and and cut me if I if I'm not answering the the your question with this I think that the the security model that Tao offers comes more from the proving side um than from the sequencing side. And why do I say so? Because um one of the things that we realized and not we but also Vitalik has said it many times is that uh all these secret circuits are are are so complex there there are so many lines of code that

is it's so improbable. it's almost close to 0% probability that there will be no bugs in the future. Uh it's very likely that within all those lines of code there will be bugs. So um that is a big issue right because you want to say that you leverage the security of Ethereum and we indeed do probably more than other rollups because of the characteristics of being a base rollup. But if uh one of the most important parts of uh of the technology which is the the secret circuits don't offer uh we we say that there's there's zero chance that there's no bugs boom all the narrative goes down right so what we did

is um is we transition right from from a CKVM there's other reasons right it's uh also on the complexity side but we transition to towards a CKVM approach and on the CKVM um we we we find some ease to to offer our users higher security levels. Why? Because there's different teams building different building different uh ways of generate proofs, different minds working on different codes. Um and and it's very unlikely now before it was very likely there were bugs. Now it's very unlikely that the same bug uh repeats across the code of all these different teams that

are building different circuits right so we leverage on a multi-proof system we use T is which is not CK um we we use SGX which is um you know Intel based uh uh uh programming uh for generating proofs but we also use CK with um succinct labs SP P1 uh and we also use CK with um with uh risk zero. So we alternate between these systems when we generate the proofs and we minimize the risk of uh of yeah of uh having the same bug across all how do you alternate? Is it kind of like um on on a block byblock uh basis or kind of

can uh can block builders uh determine what they use or what are the parameters? So it has been changing right because when we started doing this the the the cost of generating the C proofs was so high that we couldn't uh make a high percentage of of blocks that were CK proven as these teams SP1 ReZero enhance and improve their technology and make it faster and cheaper. we are trying to make mandatory um a higher percentage of blocks to be CK proven. I it sucks that I don't have the exact number now but um I think

around 20% of the blocks proposed on Tao are CK proven now and we make it mandatory like from time to time there is a blog that needs to be CK proven with our proposer uh the one that Tao runs we we now we propose 100% of of the of the blog we prove 100% of the block with CK but we we don't make the community proposers um we can we don't make them do that because it's it's it's really expensive. If there's kind of like say 20% of the blocks that are ZK proven, um does that give me any guarantees for the blocks in between that are not zk proven? So kind of like if there's a zk proven block after mine that's not zk proven. Does does that kind of confer the security

guarantees of the ZK proven block to me just because kind of like they were buil they were able to kind of build the ZK proof on that block that kind of is downstream from my block. That's correct. you have less security guarantees on all those other blocks because all those other blocks are using TES are using SDX which is also secure I would say more maybe more or not more trusted execution environment right so kind of like it's it's a hardware kind of things where kind of you you trust in Intel that kind of like that exactly exactly but kind of but kind of like if I'm on one of the TE blocks does the ZK proof of the block after mine that that does that also implicitly validate my

block and kind of give that CK I don't think so I don't think I'm not entirely sure but I don't think so though um I want to mention that uh yes it's not ideal security wise right now uh it's better than others but worse than what than what we where we're trying to get but this is training wheels for us our goal is to be a full full C actually we need it for being a stage one. So, so that's that's the beauty of the pressure that layer 2 bits layer 2 bit puts on on all of us. This is training wheels for us. We are waiting for um and they're doing a fantastic job. We're waiting for all these Cik VMs to to improve to make it faster, cheaper, so we can leverage

on them and actually offer 100% CK CK coverage. But we cannot do that as of today. Okay. But then kind of every um on every block that kind of gets um that that a validator kind of um puts forward kind of they they can decide the proving system and um the compensation for them also doesn't vary. So kind of like if if I kind of do the more honorous um kind of ZK proof I'm not rewarded more than kind of if I just do a TE proof. So you raised a very a very a very nice topic.

The way that uh uh proverbs are rewarded is is kind of like a free market. And this is interesting because some games theory play here um the proposer will get the fees from the users and and they need to get enough fees for enough fees for doing two things. One of them for the proposed block on layer one and two for paying the prover. uh the prover will put a price on the on the block. So so so they compete with each other, right? Um sometimes they can be offchain uh agreements, we don't know. Or even the same proposer can choose himself as a prover, which happens a lot of times

and that makes sense of course. Um but um I'm not uh 100% aware of how the price varies but I think it's common sense to think that when when it comes to to generate a sticky proof which is more expensive the the the price the proposer need to pay to the prover will be will be higher right because the prover need need to make a profit um but is it a double-sided marketplace or kind of like as a customer say as as a user who's transmitting a transaction or probably more likely um a DAP developer who kind of um deals with transactions from their users. Can I say I want to be on Tao but I only want to be included in ZK blocks? No. No. That that that can

that cannot happen. But it doesn't change the fee for the user. They like like the how how the fee is quoted. Um because you know because when transactions sit on the main pool they can be picked up by by whoever. There's also a competition right for that. So, so users don't need to to change the fee that they pay every time. I mean, I use Tao consistently and I don't see spikes in the same day. So, it doesn't affect the user to be honest to be honest. Okay. Um, one thing we haven't talked about at all, um, kind of like in in a world where kind of you don't have a centralized sequencer, um, you also don't have, um, a good way of dealing

with, um, ME extraction anymore. So, kind of what what what's your philosophy on that at Tao? So, we don't extract the ME ourselves, which is one of the drawbacks that I mentioned in the beginning of the interview. Um the layer one validators are the one that captures the me. The philosophy for us is um well first of all we we knew this when we decided to go this route. Uh we knew it was going to be harder for us to make a profit. Um which is fine honestly. Um I don't think like Ethereum itself they don't they don't make profit

like uh like base for example right and they are the strongest chain and we want to replicate Ethereum also even also on the business model even though it's a little bit more solid on our side because we capture the base fee right but all the MEB um we give it up we give it to the layer one validators and we're fine with that because there is this conversation on crypto Twitter of should we tax the layer twos um no Please let's not let's not uh make human interference in a free market, right? If the if the layer one offer better services like better DA, better settlement, better everything, they then then then all the rollups will decide not to go to LDAS.

They will they will happily pay Ethereum for that. We don't need to tax them. In our case, we already paying the tax via this uh sacrificing sacrificing this me and we are also happy with that because we help Ethereum be more valuable, right? We make val we kind of give another source of revenue for validators and again I mentioned um like 10 20 minutes ago that uh that not many validators are chiming in. This is going to change and we will get to that with base preconfirmations. We will give another source of revenue to validators and that also makes being a validator more you know profitable which can very indirectly impact right uh the assets.

So we are happy with the trade-off that we're making here. Okay. Um we're almost at time. So maybe um be before we kind of um uh we before we round up tell us about the Tao ecosystem. So what kind of deps are on Tao? Um and um what how do you see that changing in the future? Okay. Do we have time to go through the base precoms as well? Base preconfirmations. Yeah. Let's let's quickly talk about the base preconfirmations and then we'll talk about the ecosystem after. Okay.

Yeah. Because I think it's it's much more exciting to be honest. So uh basically confirmations uh because I think I've been introducing him this uh uh for a lot of time. I will be very quick. Okay. Basically confirmations what they make is that um all these problems that I mentioned of user experience on tao you have to wait a lot proposals don't make much money. We have to propose blocks like a bot blah blah blah. All of that get fixed with base preconfirmations. What bas base preconfirmations does they connect layer one validators with users in a way that those layer one validators make a promise to the user that they will include their transaction in a future

block. Okay, this is what happens in centralized sequencers all the time. But this is done in a decentralized manner and and and it's very cool, right? for the user. What it means is that they get they get instant confirmation of the transaction which means Tao don't need to to push empty blocks every 30 seconds anymore to guarantee a decent UX. The U the UX is guaranteed already. So all the blocks that will be proposed will be profitable by design because we can wait uh for them to be to be proposed which will be a call to action for layer 1 validators. That's one thing right. So it fixes the UX uh for for on base rollups which is very important but it does it in a decentralized manner and

how because every layer 1 validator can offer himself as a or or itself as a as a preconfirmer. They will get a a small small tip from the user. um they will need to put a bond like a like a stake in a restake model like a layer style and if they if they don't fulfill the promise to include the transaction in a future block they will get slast. So the beauty of it is that it is economically secure. It's not like a like a centralized pre-confirmation system like it's a it's like a trust me bro uh um security guarantee. in this case is

economically secure and this way we are achieving like uh again this sounds super scammy uh but uh it's actually true right we are achieving like the like the decentralization the security and the throughput in a decentralized manner right again I hate to say to talk about the trma because it sounds super scammy but it's actually true right so this is coming in in in less than one month Justin Drake Foundation shout out to because they have been leading this effort of course alongside Taico because this is uh useful for base rollups essentially and uh yeah we're super excited about it because it fixes most of the problems that we have coming in less than the less than less than a

month. It's super nice. Yeah, looking forward to that. Um maybe to come back to my ecosystem question. So kind of like talk let's kind of round it up by talking about the um the sort of DAPs that live on top of Tao. Okay, we have around 140 live dabs on Tao, which is uh not much but not bad considering how hard it's been to grow the ecosystem with the current UX. Most of them I would say are defi and gaming gaming because I mentioned before that gaming may not be the best suited for base rollup but even though we we we are huge believers in gaming and we believe that they will be suited for our

rollups or any others and they provide the the largest bunch of transactions. So we have a bunch of games uh that are quite good and we have a bunch of um of new ones coming uh on D5 we have everything you can imagine of course uh dexes lending and borrowing like all the all the basics we don't have strong perexes reason being without a good UX you you cannot have a per dex you can lose a lot of money if you have to wait one minute to get your transaction through. So we're going to push harder on on on perex moving forward. And to finish up on ecosystem, if I have 30 seconds more, I'd like to mention yeah,

I'd like to mention something that is called Tao Takeoff. Tao takeoff is like uh our biggest bet on ecosystem growth moving forward. We have realized that even the the biggest ecosystems, the biggest chains that have like thousands of applications, you know, at the end of the day, the users don't use a thousand applications. They use 20, 30 or even 50 of them in in the best cases, right? So, what we're trying to do here with Tao take off is we are selecting a few heroes on Tao and we accelerate him very very very very heavily in every manner that you can imagine. I'm I'm not going to get into details but we try to make heroes in our ecosystem. What we ask in return is that uh when they launch their

token, we usually take um we usually select like earlier stage uh uh projects. when they launch on Tao, they need to launch their token natively on Tao, of course, so that we generate um like organic TDL without having to spend so much on incentives, which is uh you know, trust me, it's like a lot of money to maintain decent levels of liquidity. And we also asked them to airdrop our community uh in several ways uh but we're one of the ways is is is to airdrop the Tao holders. So we're trying to make also um a business use case for for for Tao through this through this Tao takeoff program and it's quite exciting because you know getting getting alignment between the holders the community and the and the

applications that are heavily accelerated by us we believe is the way to grow an ecosystem is what we're going to do moving forward. Super cool. So where can we send people who are interested in Tao or building on Tao or um becoming a uh DAB user on Tao? Okay, Tao. XYC is the website. Uh the Twitter is especially I would send it to the website because you have everything you need to know there. Twitter is Tao XYZ as well. I was checking um all the updates uh happened there and uh yeah I think those and we also have a YouTube channel where we post some videos of the last base rollup summit that we did in

in San Francisco. We're doing another one in cans with uh a lot of good names coming. So you can also go to our YouTube channel and and check out and our discord and our discord of course. Cool. Thank you so much for coming on Drain. It's been a pleasure. Yeah, thanks a lot. It's been a pleasure for me too.