On The Brink with Castle Island - Jill Gunter (Espresso) on Helping Rollups Decentralize (EP.492)

Episode Date: December 27, 2023

In this episode, we sit down with Jill Gunter, co-founder of Espresso, to discuss: Espresso's original mission and the challenges of building a privacy solution Shifting direction to focus on building infrastructure to help L2 rollups decentralize The role of a sequencer in a rollup and how Espresso intends to decentralize it Trading off centralized sequencer revenue for decentralization and permissionlessness Visit Espresso for more.    

Welcome back to another episode of On the Brink. This is Ria from Castle Island, and in this episode, I'm joined by Jill Gunter, who is a co-founder of Espresso. In this conversation, Jill and I talk about where Espresso started and its decision to build a shared sequencing layer for a growing universe of Layer 2 roll-ups, the difficulty in garnering user demand for privacy, which is what Espresso was previously focused on. The role of a centralized sequencer today and how roll-ups can progressively graduate to increasing levels of decentralization and censorship resistance by leveraging a shared sequencer, which Espresso is building, the conflicting financial incentives of operating a centralized sequencer today versus the need to decentralize

to reduce trust assumptions and more. This is a media conversation, but I'm excited for you to tune in and hear Jill explain a lot of complex topics and her thoughts on the evolution of the layer to roll-up-centric future. So here's my conversation with Jill. Matt Walsh and Nick Carter are partners at Castle Island Ventures. All of these expressed by them or the guests on this podcast are solely their opinions and do not reflect the opinions of Castle Island Ventures. Guests and host may maintain positions in the assets discussed in this podcast.

You should not treat any opinion expressed by anyone on this podcast as a specific inducement to make a particular investment or follow a particular strategy, but only as an expression of their personal opinion. This podcast is for informational purposes only. Brought down by bad mortgage investments, Lehman, which has 25,000 employees, will be liquidated. The federal government loans American International Group, AIG, $85 billion. This is a different kind of market, and the Fed is asleep. The federal government is stepping it to stabilize Fannie Mae and Freddie Mac, the two mortgage giants that have been threatened by the housing crisis. The Bank of England has pumped 75 billion pounds more to British.

and zailing economy with a new round of quantitative easing. You print a couple trillion dollars and all of a sudden people start to worry. So out of this worry, we have something called the Bitcoin. Hello, everyone. Welcome back to another episode of On the Brink. Today I'm joined by Jill Carlson. Jill Gunter, formerly Jill Carlson, co-founder of espresso. Jill, thank you so much for being on.

It's awesome to have you. Thank you so much. Yeah, the little rebrand that I did when I got married trips people up all the time. You're just so notably known as Jill Carlson. I'm still getting used to it. That happens when you pivot and do a rebrand. Yeah, absolutely. I love it, though. Jill, I think you'd be helpful just for listeners. If we could start off with a quick intro, tell us a little bit about yourself. Obviously, you were one of the first people I came to know in the crypto space when I joined the industry.

and you've had an incredible prolific career and now you're at Espresso. But I think just so that listeners are on the same page as me, it would be helpful to understand your journey to Espresso. Yeah, absolutely. Happy to share. It's funny to hear myself described as someone who's been early to crypto or been in the space for a long time because I was having a conversation recently with someone where I was remembering when I was first getting into it.

this was back in 2015, 2016, starting to work in crypto full-time. I felt like I was really late. So if you're listening to this and you feel like you're late to crypto, you're not. The party is still just getting started. But I came into crypto after a stint on Wall Street. I worked as a bond trader at Goldman Sachs for a few years right out of college. And I was specifically focused on Latin American sovereign debt and derivatives. And that exposed me to a lot of folks in a lot of places who really actually needed crypto. So one of my first introductions to Bitcoin was through a colleague of mine who was working in Argentina and they were undergoing capital controls, and big sovereign debt restructuring and just high inflation rates, a lot of economic chaos.

and he started using Bitcoin to get his money out of the Argentine peso, and he started selling me on it. From there, I fell down the proverbial rabbit hole and pivoted my whole career, ended up moving out to Silicon Valley, to take a job at a startup called Chain, which was one of these startups that was doing blockchain for banks back in that era of 2016. And Chain was ultimately acquired by Stellar, the cryptocurrency Protocol. And from there, I pivoted to start working more on open protocols and systems. I had the pleasure of being an investor in or an advisor to a whole bunch of the early L-1s. So projects like Algarand, Zcash, always working with them on the go-to-market side and partnerships

and just trying to figure out who's actually going to use all of these systems and projects and products that we as an industry are developing. So that's been a little bit of my journey. I've spent a lot of time, again, just thinking about who the users are, what the use cases should be, what use cases make sense for this new technology. I've spent a lot of time revisiting my roots, given that I got started in this space and came into it through the perspective of people in places like Argentina, experiencing high inflation. I've spent a lot of time doing research, user research, and also more academic research

on that side of the world. And that has brought me to today working on the infrastructure side. I've spent, again, most of the last 10 years working more on users and applications. And espresso today is very focused on a different area, but I'm happy to get into the why of that and how that came about as well. Absolutely. Yeah, I really appreciate you sharing that background. And that's definitely a similar, I guess, development that brought me into the crypto industry, just seeing the potential for it to significantly, hopefully, improve the optionality that people have in emerging markets. And at that point in time, back in 2015 through 2017, even, and until maybe the last two or

three years, I thought that Bitcoin would be the tool. It continues to be a tool, but it's awesome also to see the massive adoption of stable coins as an incremental tool and for different incremental use cases. So it's great to see that some of the reasons that we were initially interested in this space are finally starting to play out now and actually create value for a significantly larger number of people. But yeah, going back to espresso, would love to get the quick TLDR on what exactly espresso is building. And I know maybe it'd be helpful also to understand what espresso was. initially building and then what realization that you guys had that led you to pivot to what

espresso is doing today? Yeah, absolutely. So I guess I'll start maybe just from my own perspective and what led me to work on the types of problems that we at espresso have worked on over time, which is that when I first was getting into this space, I don't have a computer science background. I'm not coming from a distributed systems or cryptography background. I just started reading Bitcoin Talk Forum and a lot of the content that was being created back five, ten years ago to try to learn about it. And I had all of these misconceptions, basically at the time, about what the technology could promise.

It felt to me reading the content that was available back then, this should be a huge improvement over the existing financial system in terms of settlement times and instant settlement times and that it should be simultaneously a much more private system because there's all of this encryption involved in it and yet also provide some level of transparency that we don't get in the traditional financial system. So I had all of these notions, obviously before I started working full time in the space about what the promise of crypto was. And I think that this is, probably a pretty universal experience, at least for those who are not coming from a deeply technical background, although I would say even for some who are. And that experience is the following,

which is basically being disabused of all of those notions of what crypto can do, at least in its current form today. And I feel like if there was a thread that I had to draw throughout the 10 years that I've been working in crypto now almost in various capacities, it would be that I am constantly on this quest to make crypto actually fulfill those promises that I thought it could going back 10 years when I first started to spend time on it. So that leads me to espresso today. We started out a couple of years ago working on the privacy side of things. And so we all had me and my co-founders and the earth.

early team that we assembled to work on this. We all had this notion that really crypto, one of the big ways that it's falling short of serving more mainstream users today is the fact that everything is transparent to everyone on chain. And that really there's a middle ground that should exist and can exist between having a fully transparent system like Ethereum and a fully private system, something like Monaro or Zcash or even tornado cash, which was obviously a popular tool on Ethereum to try to provide more fully fledged privacy. And so we spent a good year and a half exploring what was possible in this middle ground. We started out actually building an L1 to enable more private transactions on chain, more programmably private transactions. And so we started

at out building everything that goes along with that, a consensus mechanism, all of the networking side that goes along with an L1. And then we were also working on this execution layer that would enable configurable or programmable privacy. We spent a good chunk of time doing user testing with that as well, talking to a variety of people, many of whom are already in crypto, many of whom are crypto curious or crypto-adjacent to try and understand what user demand and what use cases might look like for that product. And I will be honest, disappointingly, we found it was really hard to find user demand amongst people who are already in crypto for privacy. And I think the reason is that most users who are in crypto today have gotten really used to having a

workaround for that and having to be creative and maybe go a little bit out of their way in order to get whatever privacy they want. But by and large, have gotten really comfortable with the fact that every trade, every transaction is going to be fully transparent on most of the chains that they're using. So that was a challenge with it for sure. And we started spending a lot of time then in the layer two scaling space for a number of reasons, one of which was that, We realized that the landscape of Ethereum and the landscape of crypto in general was changing. And new capabilities were being unlocked by the development of these layer two scaling solutions that were going to change the game in terms of what user experience could look like.

And another reason why we started spending time in that space was because a lot of the demand that we were seeing, actually, for privacy was coming from major L2s, who were all, of course, competing with each other for mindshare, for users. Again, this is going back over a year ago now, so many of them were in a different state of maturity at the time. We started having conversations with a lot of these teams as they were looking to differentiate from each other, and they all viewed privacy as a possible differentiator. And the time that we spent with these L2 scaling solutions led us actually to a fairly strong conviction, fairly strong conclusion about where there was real demand and where there was a real need to be filled in the market. And the conclusion that we drew

from spending the time with them was that all of these roll-ups out there, roll-ups being these layer two scaling solutions, were making huge strides in terms of delivering on the user experience that they'd been promising for years. But there was a lot on their roadmaps in terms of decentralization, in terms of interoperability, and things like this. And Looking back at what we've built so far, really of an L1, we realize that a lot of it could be repurposed to serve the needs of roll-ups. Not necessarily on the privacy side, as it's turned out, not really on the privacy side. We've pivoted away from that. But what we're building instead, fundamentally, is infrastructure to help roll-ups continue to scale while being decentralized,

as decentralized as decentralized as possible, and also allowing them to interoperate with each other. So that is a long-winded version of the journey that we've been on, but hopefully valuable for people to hear and understand how we've evolved. Yeah, absolutely. I think it's helpful to get the honest truth about whether people really care about privacy today from people who have worked on it and have done the user research. It's unfortunate. I completely agree that there isn't more demand for privacy on chain. But yeah, to your point, it's like a conflicting. property, right? When I know a lot of cases, people talk about the transparency and accountability that public blockchain networks facilitate. But hopefully at some point with the use of things, with the use of zero knowledge cryptography, will have the ability to have both. I think that ultimately there will need to be a Trojan horse that brings in privacy where people don't really know that's the property that they're optimizing for, but they get it in addition to something else.

And I think a lot of projects that are focused on privacy are doing that really well. I think Aztec does this really well where part of what they're marketing to users is the ability to get cheaper transactions. Astech is a layer to roll up and therefore scaling solution as well. And so I think that's spot on. And there are a lot of projects, again, who I am very bullish on working in the privacy space. And I hope that they make it happen because I do think that it should be the future of the way that crypto. works, but not our request at the moment at the spread row anymore. I think you guys definitely the refocus couldn't have come at a better time. If you look at some of the analytics around transactions taking place on L2s, in aggregate, it's significantly higher than L1s. It seems like the roll-up centric future. It's actually coming to fruition. I think it's like 4x right now is the number of transactions on Ethereum L2s versus on the Ethereum base chain, which flew my mind when I learned that. Yeah, it's crazy. I was looking at the charts earlier today. And I think earlier this year, people falsely assumed in some instances that on-chain activity has gone down because they were

just looking at L1. But it's like when you look at it in aggregate, it's really taken off. So going back to espresso, and I think espresso in the context of some of the challenge is that L2s on Ethereum, as well as maybe other networks in the future, are facing today. And maybe it's not really a challenge more than this is the phase that they're at in their roadmap towards a future state where everything is permissionless, censorship resistant, decentralized, what have you. But I guess, yeah, just taking a step back, what is the steady state today of the majority of roll-ups on Ethereum?

Yeah, absolutely. So the majority of roll-ups on Ethereum today still have some training wheels on to some degree. I think that's literally what Vitolik called it in this post that he wrote up about the progress and the evolution of roll-offs and what it can look like, but it should look like. And he outlines these three different phases

of sort of levels of decentralization, levels of therefore censorship resistance that roll-ups can offer users. And I think it's a really valuable framework because I think all too often it's very opaque to users what the risks are, what the trust assumptions are when they're using blockchains in general. And shout out to L2Bit.com. They've done a really great job. Maybe they even did it before the Vatolic Post, but they've done a really great job of implementing a very easy way. for users to look up and research and understand the different phases that each of these roll-ups are at. And so there's one or two roll-ups that are pretty far along in their

roadmaps that would be qualified as phase one roll-ups. And then most of the others on Ethereum today have still sort of full training wheels on. We can talk more about what specifically that means, but in short, it means that users are still having to make some fairly aggressive trust assumptions in certain circumstances in certain cases when they're using roll-ups. One of the ways in which they're having to make trust assumptions on pretty much all roll-ups today is when it comes to the sequencer component of the roll-up. And so maybe it would be helpful, I don't know, you tell me if we took a step back and went through the different components of the roll-ups, of stack. I think that would be super helpful. It's not really something that we've gone into in depth

on the podcast. So I think, yeah, taking a step back and understanding, even just starting with the sequencer and the different roles that it performs and I guess what the risks are of having a centralized sequencer would be helpful. Yeah, absolutely. So roll-ups fundamentally, roll-ups for L-2s, I'm going to use these phrases interchangeably here. Roll-ups fundamentally, move computation off of the layer one. And that is how scaling gets achieved through them, more or less. Now, when you move computation off of Ethereum or off of the layer one, you have to introduce whole other architecture, a second layer of architecture, if you will, in order to enable transactions to still be submitted and processed, and then ultimately

gaining finality from the layer one in this case. Most of what we're talking about is Ethereum. And so you wind up with all of these different components of this second layer of architecture that is still ultimately feeding back into the L1. And so you have the execution engine or the execution rules of this layer two. Most of the major L2s are aspiring to be or coming close to being fully EVM equivalent to try to make it as easy as possible for developers to be able to take their existing app on Ethereum and just deploy it on the roll-up without having to make many changes. So that's the VM, right? That's, again, this execution engine. Then you also have a set of on-chain contracts that are taking and reading and in some cases

storing the data of what's going on within. in this, again, second layer of architecture. And when I say on-chain contracts here, these are contracts that, of course, live on the L-1 itself, and thereby enable the L-1 Ethereum to be used as the settlement layer for ultimate finality, meaning that users in the ideal case, in the fully mature case of this second layer of architecture, should still be making the same trust assumptions

as they're making when they use the Ethereum L1 itself. Then you have data availability as concept or a component, if you will. In most cases, this is being posted to Ethereum still as the data availability layer. Again, I think that a lot of roll-ups are very opinionated about this. A lot of roll-up teams are very opinionated about this right now because they want to stay as close to Ethereum as possible. That's very expensive for roll-ups in terms of, percent of fees, most roll-ups are paying 70, even 80 percent of their fees back to Ethereum

in order to post the data or the data availability piece of this. Quick question on that piece. Do you know what the margin is for roll-ups if you think about the total cost that they're paying with 70 to 80 percent being data availability on Ethereum specifically versus the revenues that they're generating? Yeah, absolutely. So the main cost for most roll-ups for almost every roll-up today is that is posting data to Ethereum for DA, as we call it, data availability. ZK roll-ups also have some additional costs in terms of proving costs. But on average, certainly the lion's share of activity going back the last couple of years

has been in optimistic roll-ups as opposed to ZK roll-ups. We can also get a little bit into the difference between these two. So if you look at historicals, it's more 25% margin in terms of what the roll-ups are being able to take. Because again, it's mostly the data availability. From what I've seen of ZK Synx data, which is sort of has the most data amongst all of the ZK roll-ups out there, they are pretty much in line with that despite having higher costs because they do have this proven component as well. So that's worth noting. But this is a huge topic of conversation amongst roll-ups is how to continue to improve that margin profile and how to be able to reduce data availability costs. And of course, Ethereum itself has some initiatives around this. The Ethereum Foundation and their researchers are working towards dank charting, which is a great word. Go look it up. I'll save that for a different episode. And then you also have Celestia and Eigen, DA. We actually have, our own data availability system that folks can opt into using as well as PresoDA. So, yeah, there's more a growing sort of competitor set for how to drive these costs down,

certainly. Yeah, that makes a lot of sense. I think it's great to see L-1s optimizing for this roll-up-centric future and really trying to make it easier and cheaper and more efficient to operate a roll-up. and kind of accepting it as the future, as the architecture that we're all moving towards. So the final component that I haven't touched on yet that is the most core component to what I'm working on is the sequencer, which we've mentioned a couple of times already. But the sequencer effectively takes the transactions that are being submitted to be processed on the roll-up

and batches them together and to a degree orders them. Now, the degree to which the sequencer is smart or clever about ordering them, that can vary depending on what your definition of sequencer is. The way that I tend to think about it is that the sequencer is not especially smart when it comes to the way that it's ordering the transactions, although it could be. But for the most part today, the way that the sequencer gets run by roll-ups is just as the centralized entity associated with a roll-up is running a server or set of servers that are performing this as an operation. And for the most part, those servers are ordering transactions just on a purely first-come-first-serve basis. Now, what exactly first-come-first-serve means is a whole Pandora's box that we can get into as well. For the most part, that's the transaction ordering policy that these roll-ups in general are playing. And then the sequencer, again, is just accepting transactions. batching them into blocks and then submitting them the next steps of processing by the roll-up VM

to be executed, and then also talking to the L-1, which of course is still being used as the settlement or finality layer. But the important thing to note here is that today, the way that these sequencers work and have been architected and operate, they're almost exclusively centralized. It's on pretty much every roll-ups roadmap to decentralize the sequencer at some point, and we can get into why. And it's also worth noting, though, that these sequencers are, and roll-ups in general, tend to be totally siloed for each other. If you think about that, that to me is one of the bigger problems here, which actually does not have much to do with decentralization, although it's not totally divorced from it. But it's just the fact that they're siloed. The beauty of Ethereum is that you have all of these applications and all of these smart

contracts that can call each other and really seamlessly interoperate in the same environment. If you talk to any on-chain gaming developer, what are you most excited about these systems? Something that always comes up is the interoperability component, not have to be in their own walled garden. Roll-ups kind of introduce walls into the infinite garden of Ethereum. And so one of the most interesting questions to me and interesting questions to us, espresso, is how can we start to break down those walls again that are popping up as

Ethereum does move towards this roll-up centric roadmap. And so to cut to the chase, what we're working on is a decentralized shared sequencer layer. We sometimes talk about it as a layer 1.5 because it's an architecture that will sit between the L2 and the L1s. We're adding layers. Indeed. And that is what we're developing is this infrastructure layer for sequencing that can be shared across multiple roll-ups, help to support interoperability, and can remove some of the trust

assumptions or risks that can come along by a roll-up operating just its own centralized sequencer. Got it. That makes a lot of sense. One question that I have, I wanted to get your thoughts on this before we dig into what a shared sequencer looks like is if you look at the revenues that some of these roll-ups with centralized sequencers, single sequencers are generating. It's non-trivial, right? It's a significant amount of revenue that they're inheriting. And so there's maybe this financial incentive to continue operating a centralized sequencer, right? But then on the flip side, you have this conflicting duty maybe to the users in Ethereum and in these roll-up ecosystems around censorship resistance, permissionlessness, and interoperability, as you mentioned.

How do you think about, and maybe in your conversations with some teams that are operating roll-ups, how are they thinking about these conflicting points of view? Yeah, absolutely. So you bring up a great point because amongst the top, I think it's just amongst the top four roll-ups today by transaction volume. They're already on track to create about 200 million in revenue on the year. That's taking the year-to-date numbers and extrapolating them out here for the last two months in November and December, which is significant. And especially in crypto land where revenue can often be all too foreign of a concept. That's huge. Now, it still pales in comparison, actually, to Ethereum, which is on track to

process about $2 billion in dollar amount in terms of fees on the year at current prices. So it's an order of magnitude difference. Even though roll-ups are doing, remember, about 4x, the number of transaction the L-1 is doing. So that tells you something, firstly, though, about roll-up cost savings, which is great for end users. But, Most roll-ups today, I think, are being quite thoughtful about how they treat these revenues and how they treat the net profit that they're making. And most roll-ups today are very realistic about the fact that is not money that's lining the pocket of any given project or any given entity, but rather is money that is being returned to the community and being returned to. the project in kind of a larger sense, right? It's not actually being recognized as revenue of some centralized organization or company, but instead is being used to fund perhaps the treasury of a

Dow that is then doling out grants either retroactively or in advance in order to fund more things on the roll-ups roadmap or in order to fund developers to come and build on the roll-up. And so I think that that's a really important starting place to understand the mindset of roll-ups because I hear this question all the time. If I'm a roll-up and I'm making extensive millions in revenue this year, why would I give that up? But it's not really being conceived of in... That's like a long-term reality. Yeah, exactly. Centralize corporate sense of it. And I think that's smart. And I think that demonstrates good foresight and also just good stewardship of the crypto ecosystem. Think about what it would look like if Ethereum was running some

centralized components via the EF. And the EF was pocketing some amount of profit from those two billion in fees that are running through it. We'd all be like, wait, what are we doing here? This doesn't actually make any sense. And so I think it demonstrates a real sense of where the value add comes from these types of protocols and the right way to be stewards of them. And also a sense of realism around the fact that, again, on it pretty much every roll-ups roadmap, there is this item of decentralizing the sequencer. And certainly once you do that, then where the fees go becomes a much more complicated, less straightforward matter. And so better to start tackling that and thinking of it as funding for the project and community as a whole,

as opposed to the way that, again, an old school corporation would conceive of it. So most of our conversations with roll-ups, this does come up, of course, of what sort of the breakdown of fees look like and how roll-up projects can make sure that they will still be getting a cut of fees that are being run through the roll-up and the sequencer and the L2 and the L-1.5 and the L-1-itself and what have you. And perhaps more than a cut, the majority, perhaps, making sure that will go back into their communities and create the right incentive structures around their projects. And that's something that we at espresso hold very sacred, actually, is to make sure that we, as the builders of this infrastructure layer that is looking to serve

roll-ups, were incentive-compatible and incentive-aligned with the roll-up projects in their communities themselves. So there's a number of ways that we think about our infrastructure layer and the nodes running espresso capturing value that we think are not at all at odds or incompatible with the way that roll-ups end up going about that for their communities. So we can get more into that. But it is a fascinating thing because until recently, I think, you haven't seen a lot of discussion of true economic models around these types of things when it comes to protocols.

And because roll-ups are in this somewhat unique position of evolving from being a bit more centralized to being very decentralized, it does bring up some interesting points of conversation like this. Yeah, absolutely. I think it'd be helpful to understand as a segue shared sequencer, or what does espresso's shared sequencer design look like? You mentioned that you're positioning espresso as an L 1.5. And then going back to this conversation around roll up sequencers today, it would also be helpful to understand what role they play in. in this future model and do they become one of the nodes on the broader espresso network or how does their function change? Yeah, absolutely. So just to start maybe with what espresso is or the

espresso architecture, you can think of it in many ways as being akin to an L1 in terms of architecture where it has its own consensus mechanism that's running. It is a proof of stake protocol. It will have a decentralized network of nodes and running the leader election process, as many L1s do, in order to determine who is the leader for a given round, who is proposing a block of transactions, et cetera. Now, importantly, because we are not a settlement layer, we get to do things differently from the design choices that say Ethereum has had to make. So whereas Ethereum has had to make, make a set of design choices that very much makes sense if you're a settlement layer, if you're being used as the last resort of settlement finality, meaning that they have had to

optimize for things like dynamic availability, wherein even if 90% of the Ethereum nodes go offline, Ethereum can still keep running. We can make a different set of assumptions and therefore meet, I think, a different set of user needs just by virtue of where we're positioning ourselves in the stack. And so we instead have made a set of choices when it comes to our consensus that can really optimize for things like scale, throughput, low latency, to get pre-confirmations, all of these things that roll-up users have come to assume are going to be part of the roll-up experience, basically. And of course, it's part of the roll-up experience today for many reasons, for many architectural choices that these roll-ups have made,

but not least of which is because they also are depending, at some point, on a centralized server to play this role. We're trying as best as possible to make sure that, again, the design choices that we're making on the consensus side support that level of scale. So we use a consensus mechanism called Hot Shot, which is based on Hot Stuff, which was developed initially actually for the Libra project and protocol that came out of Facebook, which gives you some inklings of the types of design choices that were valuing. Obviously, that was very much trying to be like for the mainstream,

for the masses and really optimized scale. We've made some changes to the hot stuff protocol to develop it in a hot shot, one of them namely being making sure that we can support more fully fledged, decentralized network of nodes as opposed to having it be more of a permission system like Libra was going to be. We have on our website, I think, a decent sort of map of what all of this architecture looks like. But again, you can think of it very much akin to an L1, again, but we get to make different design choices. And another one of those is actually the fact that we don't have to maintain state. And so we can be a much lighter weight system than an L1 would be.

And again, this system sits between the L2 itself and the L1, which the sequencer is posting to. And again, importantly, the sequencer is also working in our case across multiple different roll-ups to be able to support not just decentralization, but also this notion of shared sequencing and interoperability. I think your second question was about why decentralization is. important or matters with the sequencer role? Yeah. I think there's another question that came up that I want to double click on as you were describing the architecture and the design choices. I think it'd be helpful to understand despite not maintaining state, how do you achieve or allow the roll-ups leveraging

the sequencer to achieve interoperability? Yeah. So there's a few different. mechanisms through which interoperability becomes easier with a shared sequencer. I'll say up front that on pretty much none of them is the shared sequencer on its own silver bullet. And so this is a place where I actually get a lot of questions or confusion from people of like, oh, so does a shared sequencer replace a bridge in this scenario or this or that? And in most cases, not necessarily, but it can very much augment what's possible. So one of the ways in which a sequencer can create better interoperability is, if you're familiar with the concept of builders in Ethereum, this is getting into

kind of the realm of proposer builder separation, where you have in Ethereum, if you think about it from the L1 perspective, you have validators who are, of course, running Ethereum consensus. But then you also have this separate class of sort of specialized actor, block builders who are assembling blocks and then submitting them to the validators who are acting as proposers to the system to propose the block to the network. And these block builders can do a variety of different things in a way that they order transactions, many of which are aligned with user incentives and user needs to make sure that users are being treated fairly or not being front run or sandwich attacked or any of these things that arise or at a minimum are creating such a

competitive market of builders that they're driving the value extraction that users have to experience down to as low a level as possible. Now, if you think about this in the realm of L2s, what does this look like? In order for builders to do what they do, and to be able to include transactions that are happening across multiple different roll-ups and execute the transactions that are happening across multiple different roll-ups. If you think about that world without a shared sequencer product, suddenly those builders are going to have to be working across N different sequencers that they're submitting to.

And their block proposal might not get picked up by all of the sequencers that they need to be submitting to, in order to guarantee interoperability and execution across all of them. If they only have to coordinate with one proposer at any given time, which is hopefully the case if we get sufficient adoption with espresso amongst roll-ups, then that is a much more straightforward and much more economically rational game for the builder to be playing. So this, I think, is a great example where, in my view, a shared sequencer is an absolute game changer in terms of what you can achieve with interoperability. But again, it's not just the shared sequencer in a vacuum. It's the shared sequencer plus these sort of builder actors

or entities that are operating the system as well. That's one good way to think about it, I'd say. Got it. Yeah, that's super helpful. And then, yeah, I think the second part of my earlier question was around what role do potentially existing roll-up sequencers have the potential to play in this new shared sequencing environment? Can they operate a node and contribute to being one of the processors on the network? Yeah, having roll-ups participate as nodes or at least as stakers delegating to two nodes on the network is very much one of our goals and very much one of the ways that we're going to seek to align incentives and have things be really compatible between espresso and the various rollout projects. And that's something that we actively talk to the

rollout projects that we're partnered with about exactly. Awesome. So over the course of this year, there have been a number of roll-ups that have launched via the OP stack optimism, which operates his own roll-up, but also has this framework and standards and tooling that make it easy for teams to launch their own roll-ups. We've seen base as an example, Zora is another example. There are a number of other examples. And optimism is trekking towards this super chain future where you have a standardized set of roll-ups. And I'm not super familiar with how they're thinking about the sequencer. So it'd be great to hear from your perspective.

What exactly is their plan as it relates to sequencing? And then is there a future in which espresso plays a role within the optimism or O.P. stack? Yeah, we certainly hope so. that's up to OP governance eventually. Once we get to main net, that's something that will be paying a lot of attention to. I love the articulation of the optimism super chain vision, which for those listeners who may not be familiar with it, it's effectively this idea that we should have seamless interoperability between roll-ups and roll-ups able to call each other

just the way that different smart contracts on Ethereum can. Optimism's articulation of the super chain vision, of course, is more OP stack-centric, as opposed to, I think, the one differentiation that I would make is to say that I think our vision for this super chain is Ethereum-wide, as opposed to just on a stack-by-stack basis. But I think that they were among the first to really articulate this as a problem within the roll-up space. And I think that they've done that very well. So we're working closely with them on a number of fronts, most particularly an RFP process that they kicked off a few months ago that we've had the privilege of working on alongside their teams at OP Labs and the Optimism Foundation,

their broader group of contributors. This is the first step towards sequencer modularity or sequencer decentralization and shared sequencing. And we've supported and led actually and supported the build-out of plug-able way to integrate a sequencer into the OP stack. That, to be clear, is not an espresso-specific thing. This component that we've been working with them on is really a more general contribution to the ecosystem than just for us. It's for anyone who wants to come in and plug in and start to experiment with and perhaps even implement into production an alternative sequencer.

from what is today the optimism, more centralized sequencer. So that's been really excited. We as a company really love making these types of broader ecosystem and open source contributions. Of course, it's also helpful to us to do this work. This is work that we would be having to do anyway as we work on an optimism integration. Our current test net that's up and running has an OP stack integration, has an OP stack roll-up running on it in partnership with Caldera, which is a roll-up as a service company that allows

sort of one-click deployment of roll-ups. And indeed, they did a one-click deployment of an OP stack roll-up on Espresso for this most recent test net of ours. So that's been really exciting. But I'll also say one of the things that I love most about my job at the moment working on the go-to-market side of espresso is that every arrangement with every different roll-up team and working relationship and what that looks like is so vastly different. So we have an amazing partnership with off-chain labs, which of course is the team behind the development of the Arbstrom L1. We're working with a number of more up-and-coming roll-up teams, teams like Spire and Opside, etc. And also working with a number of roll-up as-a-service teams. And again, each of them has very

different priorities and very different approaches to the space. We have an amazing more technical research collaboration going with off-chain labs as an example, working on the builder side of the world that we touched on a few minutes ago. That's again, one really cool part of this is that there isn't a go-to-market playbook for how to work with roll-ups and how to work as this sort of new layer of infrastructure. We at espresso are getting to make that up as we go along here, but it seems to be going okay so far. It's a very interesting and exciting set of opportunities and challenges to work on. I'm sure there's never a dull moment. Yeah, and totally. There's always a new roll-up cropping up, which is great. That to me has been, just from an

entrepreneurship perspective, a huge lesson of just launching something into a growing market, which we've stumbled into here. If you'd asked me a year ago how many roll-ups I thought would exist or be in the process of being built in the fall of 2023, I probably would have said maybe 10. And it's hundreds, which is really exciting. And I think that it's a really exciting new vision for what the future of Ethereum ends up looking like. Definitely. That's a really good transition into another question I had around just the general trends that you're seeing in the roll-up or L2 landscape across maybe teams operating general purpose roll-ups, teams launching app-specific roll-ups, operating or making the decision to use data availability on Ethereum versus Celestia

versus maybe some of the other solutions like those released by EigenLair and Polygon. What is this status quo of the landscape today across some of these decisions? I think what's cool is that we're seeing roll-ups really mature, actually, in terms of how they approach their projects in a way as a business, which I know runs in the face of what I said a few minutes ago about roll-ups being very mindful of the fact that the quote-unquote revenue that they're getting from their centralized sequencers today is getting returned to the community. But nonetheless, I would say roll-up teams are approaching their projects as business. And they are mindful of things like margin or things like user experience that formerly, I would say,

in crypto, we've gone through phases where almost ideology trumps everything else. But I would say the big trend that I've seen amongst roll-ups, even just over the last six months, is just a real pragmatism in terms of wanting to ensure that the product makes sense for users, that users are getting their needs met, that there's a realism amongst roll-up teams, that they are in a war with each other in many ways for developers, for end users, and doing whatever it takes to win those users, win those transactions onto their chains. And so you are seeing a lot more roll-up interest in alternative DA layers, not just Ethereum. You're seeing a lot more interest in ways that they can not compromise on their roadmaps around

decentralization while still maintaining scale. And you're also seeing a lot of focus from these teams in terms of, again, interoperability because they view that as a key go-to-market lever that they can pull. and a key value proposition they can offer to developers and users. I think that the idea of having shared liquidity across multiple roll-ups to a lot of these teams is something of a holy grail. That's really heartening and exciting to me as someone who spent, again, most of her career focused on the application side of the world

and not just the infrastructure, is these teams really taking that lens of, okay, how do we design these systems, not just so that they feel really clean and ideologically aligned to talk about, but so that they're actually serving a purpose and everything is very much thought through from that user perspective. Yeah, absolutely. I think that is such an important call out. And to your point, I think these operators do have to put on that kind of business hat at some point, right? It's not just about the tech, unfortunately, is something that we've seen being in the space for as much time as we have. I saw Victor Yunnan from Coinbase tweet something today about, I think it's a hot take.

There's nuances to it, but the best and most successful role-up teams are those that have really strong B-D teams and have someone on the other side to pick up the phone when there is like a B-D or partnership or user-related question. Well, to the extent that we're having any success on that, at Espresso, full credit, goes to the team around me. Great BD team and go-to-market team that we have. It's such an important function. Absolutely. And if I think about the L-1s that we saw develop over the course of the last cycle, sort of 2017-2018 era, I think it was often ignored until too late.

And then at that point, it just became this sort of approach of just three, throwing money or throwing tokens at the problem of trying to attract developers. Maybe with the exception of Solana, actually, I think Solana deserves a lot of credit for having really thought through what the right balance of tradeoffs would be in order to more organically attract developers. But I think that it's something that, again, is refreshing to see for this cycle's infrastructure projects that BD, go to market, who are the users? It's not being left to be an afterthought this time around. Absolutely. It's so much more intentional, which, yeah, makes me really optimistic about the space and the type of talent that we need

to really reach mainstream audiences, too. So as a last question, Jill, I know we're coming up on time here, but I'd love to understand here what's exciting you the most as it relates to unique app or in for opportunities and innovations that we can unlock by this greater amount of transaction throughput that were in improvements and cost, speed, user experience that we're facilitating through just pushing forward this L2-centric future. This is maybe going to be such a boring answer, but the thing that I'm excited about, the application that I'm excited about being enabled by this latest round of infrastructure is actually just payments.

Preach. I totally resonate with that. Yeah, you come from the Bitcoin mindset as well, Rhea, but I very much do. There's probably not a month that goes by that I don't revisit the Bitcoin white paper at some point. And I feel like it laid out such a clear vision for why payments on blockchains make sense. And then if you just look at your own day-to-day life, sending a cross-border wire

transaction. You're just like, for the love of God, there's got to be a better way. Even sending a domestic wire transaction is insanely complicated and expensive and slow. And then you compare that to just sending USDC on Ethereum. But I think that there's so much in the world of payments that has been held back by these unpredictable fees on blockchains, and even through stable coins and all of this kind of thing, that I just, I think that this is going to be the time. I think that this is going to be the unlock. And that's something that, yeah, I try to keep in mind as I'm getting lost in the infrastructure weeds is what are we enabling here? And I just always tend to come back to payments. But there's a lot of other spaces that I'm excited about in terms of on-chain social and gaming and things like this. But I feel if we as an industry could crack payments, never again will we have to deal with

the mainstream media saying things like, where are the use cases? What is it good for? Are those painful Thanksgiving dinner conversations that are coming up for all of us? Where our parents ask us. I need to check the data for this year, but I'm curious to see where we stand as it relates to stablecoin transactions facilitated across different L1. and L2s and how that compares to total payment volume on traditional card networks or payment processors or whatnot. I'm assuming it's grown to spike this bear market that we're in. And yeah, I completely agree with you. I think it's one of those no-brainer use cases for this technology.

It just makes so much sense that this activity would transition to being on chain. So in complete agreement. Let me know when you dig up those numbers. I have a feeling I'll be woefully low relative to card payments that I think we're getting there. Yeah, definitely. Jill, it's been such a pleasure to have you on. Thank you so much for all your time and for walking us through some of these more technical concepts. I really appreciate it. My pleasure. Yeah, thank you so much. And I'll just say as a final note to anyone listening who's not deep in the infrastructure weeds. Don't be intimidated. I've had probably on average 36 tabs open with these various concepts and acronyms and so forth up on my screen for the last two years, catching up on all

of this and keep at it and feel free to tweet at me or reach out with any questions as you continue on your journey. Or if you're a roll-up project looking for a shared sequence for solution equally, hit me up. But thank you so much, Rhea, for having me on. This is a pleasure. Thank you so much, Jill. Thanks for listening to another episode of On the Brink with Castle Island. To find out more about Castle Island, visit castle island.V.C. To listen to all of our podcast episodes, please go to On the brink dashpodcast.com or just click on the tab in our website. Thanks for listening.