The Next Frontier: Trustless Compute w/ Tim, Wolfgang and Rami (RISC Zero)

**Speaker A:** Hello and welcome back to the Strange Water podcast. Thank you once again for tuning in. When I entered this space in 2021, I saw through the lens of finance. After my time running the treasury desk and then controlling a 10 figure budget at one of the world's largest companies, I was convinced that Ethereum was here to revolutionize the way finance is done. Now, before I go on, let me be clear. I still think that Ethereum is going to revolutionize finance. I have so many stories about how the actual plumbing of the global financial system just doesn't work. Crypto is better, Ethereum is better. But over the last 2ish years, my views have evolved significantly. While I still think that Ethereum is going to change finance, I don't think that that's the purpose of Ethereum. In fact, I am becoming more and more convinced that Defi isn't going to be a particularly important part of the world computer. In fact, I think that Defi has a lot more in common with NFTs than it does with global finance these days. I believe that the purpose of Ethereum is not trustless finance, it's trustless compute. Just like the purpose of computers is in finance, computers are used for so much more big data collection and analysis, human connection and social networking, content creation, scientific modeling and research. Playing games. I could literally go on forever. And while we can all think of ways to add tokens and make these actions fit into our current conception of Ethereum, there's no reason to think that everything needs to be financialized and speculated on. Now, all of this may sound interesting, but it's super high level. Are you having trouble grasping what I'm talking about when I say trustless compute? Well, today I'm pleased to bring you Tim Carstens, Wolfgang Wells and Rami Khalil from RISC 0. Over the next hour, we are not only going to unequivocally spell out what trustless Compute is, but you're about to learn that it is ready right here, right now. And Ethereum, crypto and computing will never be the same. One more thing before we begin. Please do not take financial advice from this or any podcast. Ethereum will change the world one day, but you can easily lose all of your money between now and then. Okay, time to bring on the folks from Risk Zero. Hold on to your seats. Foreign. Welcome to the pod. Thank you so much for joining. **Speaker B:** Yeah, thanks for having us. **Speaker C:** Thank you. **Speaker D:** Yeah, right on. **Speaker A:** So this is the first time I've ever done a three person interview and normally I like to start the Pod with like, just the background and how you found crypto. So maybe we can just kind of like, keep it tight. But I would love to hear for all three of you, like, how did you find crypto? And like, what was your moment where you realized, like, oh, this isn't just Internet money for buying drugs, but this is something like, worth building a career, if not around adjacent to. So maybe Tim. **Speaker D:** Yeah, so I've had. Crypto has been a close and dear friend for a while, I would say, but definitely with its highs and lows. I first came into it, so I was in math grad school when Satoshi published his paper on bitcoin. And I also came from kind of that cypherpunk kind of scene. And so I saw that paper pretty early on, and for me, right away it was just like the nerd in me was very excited. I said, this will kick off a whole bunch of new ideas, new algorithms, new Internet experiences. I didn't, however, think that people would more or less just take that design and run with it and that 15 years later, you know, these tokens be worth tens of thousands of dollars. I don't think many people saw that. So I kind of hung out in the forums for several years. And then around 2012, one of my friends and I, we got into mining. We did that for a while. We got goxed and then, you know, more or less, actually around the time that people started the early work on the design for Ethereum, I kind of decided for myself. I said, okay, do I want to stay in this for a while and really go with it for, you know, all the way, or, you know, I was still pretty young. Do I want to check out some other things? So I actually got out of the space for a while and then, I know, still holding, but, you know, just, you know, other kinds of research, formal methods, other kind of fun things. And then I came back into the space at the start of last year when one of our colleagues at Riskro, a guy named Jeremy Brussel, reached out to me and asked if I had heard of zero knowledge proofs. **Speaker A:** Oh, all right. So that's going to be the meat and potatoes of our conversation. So I'll pause you right there and talk about, like, we'll get back to what ZK is and basically what it is outside of the realm of crypto. Twitter. So super excited. Thank you, Wolfgang. How did you find the space? **Speaker B:** I guess I was a math PhD student when I stumbled across the bitcoin white paper. And to be honest, I was not impressed I was kind of missing the math, the cryptography, so it didn't catch on. So I basically left it there for a while and then like it was like 2017, 18 during the hype, I was looking for a new job and I said, man, now is a good time to brush up your GitHub portfolio. What interesting projects are there? So then I really started doing like community development for some projects and I also ended up then working for, it's called iota, so that very strong research focus which I really liked being a researcher myself. And they were building basically their own layer one, which was a very, very interesting endeavor to do. And then I got into this even more really had you had to understand how the consensus system works. That is actually just more or much more than, I don't know, just the white paper. What I thought a few years earlier. So much more or also like smart contracts cryptography. I got a lot into cryptography and that's also how I then got into like zero knowledge proofs being basically. Yeah, the, I don't know, the king of cryptography or the third, I thought at least. So that's how I ended up at risk zero then. **Speaker A:** Oh God. I'm really afraid that Rami is about to tell us that he was a like actual math PhD or studying math too. And then I'm going to be the dumbest person on this call. But at the risk of being dumb, Rami, how did you find crypto? **Speaker C:** No, Well, I mean, I do have a bit of an academic angle getting into crypto, but I started maybe a bit earlier because towards the end of my master's we had this lecture on selfish mining in bitcoin and that was kind of fascinating and that got me interested so I started looking at payment channels. I made this scheme for rebalancing payment channels and then eventually started a PhD around blockchain. **Speaker A:** So sorry, I did not mean to belittle at all the academics. Like, I mean, just so you guys know, real briefly, I went Stanford from 2009 to 2013, studied computer science. I took intro to number theory from Professor Dan Bone. I read the bitcoin white paper in 2011 and the entire time I just thought like, I mean, kind of cool, I guess, but like, not relevant, I don't care. And I left the space until 2021. So yeah, you know, I think like we all. And I think too like during those 10 years, like I worked at like the largest beer company in the world. I became like the cash flow manager and I was the director of finance for Bud Light. So I think we take all of our experiences and that informs what we do in this space. But it is pretty wacky how we all end ourselves up here. And I'm glad none of us found it through buying NFTs. So I think before we really dig into this conversation, I think it's helpful to put into context what, what first of all, like, what is risk zero? And like why risk zero? So maybe I'll start with like kind of a theory I have on cryptography and then let you guys kind of like build off that and then tell us what risk zero is. But I, I've become like super convinced that like we created cryptography in World War I and World War II to like encrypt messages for war, right? And then for like I don't know, 80 years, we got better and better and better and better that. But like we thought the purpose of this science was encryption. And then in 2008 Satoshi came along and said like, hey, there's not one use case for this, there's at least two. And for the last, I don't know, 15 years we've been like chasing that thought, like oh, cryptography is for one of two things. And I think all of us on this call and like all of us who understand cryptography realize like if it's more than one, it's not two, it's end use cases. We just need to figure out the other end use cases. And like I've got some thoughts on like what the purpose of cryptography or specifically zero knowledge cryptography is. But like that's kind of how I understand where we are in time. And so one, I just want to hear like your guys's reactions to that and to agreeing or not, like how do you put into context like what you're building and like what brought you together for risk zero? **Speaker B:** Maybe I start with this general stuff about cryptography because I 100% agree on this. Like for example, a lot of stuff like the theory, if you look at the papers or whatever, they're quite old. And then really crypto gave all of this a huge boost because suddenly there was the motivation, suddenly there was the use case, suddenly there was funding like outside of universities. So I think that is one of the, in my opinion one of the best things that came out of all this crypto hypes. And for example, I don't know, there's some theory about like threshold or multi signature schemes. They're quite old. But why would you. There's no real use case but now suddenly, if you have like some, some, some, I don't know, tokens, and you have a board of directors and you wanted like a 3 out of 5 signature, suddenly, suddenly you need it. And it needs to be relevant and it needs to be secure in the real world and not only in certain theoretic aspects. And so I really think that this gave, the theory was there, but how it can be applied, how it can be made fast, how it can be maybe optimized in hardware, whatever, that really was driven by the whole crypto hype. And I'm honestly excited. What happens there next in the future if this continues? **Speaker A:** Yeah, for me, when I decided to figure out what was going on here, and I read like all the practical Byzantine fault tolerance papers, and I'm like, they're all written in the 70s, like, what's happening? **Speaker D:** Yeah, well, Zero Knowledge is a good example of this. The theoretical framework for it was already known in the 1980s. And it's exactly one of these things where the theorists could envision somewhere off past the horizon, a bunch of hypothetical scenarios. But they were really fairly contrived because at this time we didn't have the Internet, we certainly didn't have an Internet with over a billion users and who knows, you know, how much economy, economics and all the other great activity that takes place on there today. And so, you know, to kind of echo what Wolfgang has said, you know, it's been really exciting for me to see a lot of the stuff that I kind of considered, like really far out, very pure math kind of ideas from computer science actually coming in and being the hot stuff that, like, is really unique in its ability to solve some of these core problems. **Speaker A:** Yeah, makes a lot of sense. So with that being said, like, what is the core problem that Risk Zero is trying to solve? And actually, I guess like, even before that, what is Risk Zero? **Speaker D:** Yeah, I'll take this one. So Risk Zero is a zero knowledge technology provider founded a little over two years ago by Jeremy Brussel, Frank Laub and Brian Retford. Our key product is an open source platform. Well, it's a library technology, whatever you want to call it, called the zkvm. So this is a gadget within which you can run programs. And the important thing is that whatever output your program produces, you also obtain a proof that that output really did come from that program. So this is a very geeky kind of thing. But if you happen to be in a space where you have a bunch of computers working together that have no good reason to trust one another, it actually winds up being super useful. **Speaker A:** When we think about this Cryptography vm, is this a full Turing complete environment which you could run anything, or do we need to think about at least what we have from the beginning as a specialized computing environment? **Speaker D:** Good question. Yeah. So the key thing about our ZKVM is that we see it as a completely general purpose computing solution. And so specifically for the builders out there, we provide an industry standard RISC V instruction set architecture together with a number of accelerator co circuits that you can invoke from there to get you higher performance for cryptography, which fun fact is exactly how regular computer processors also work and do those same operations. So we really would like to see it as something that has a broad impact across Internet applications and really all sorts of decentralized applications. But of course, the space that we're most excited about is blockchain, because blockchain understands what this stuff is and is already very excited about it. For sure. **Speaker A:** It makes a lot of sense. So the base kind of product or concept Is this cryptography VM or sorry, what did you call it? **Speaker D:** ZK VM. **Speaker A:** ZK VM? There's too many VMs I need to keep track of these days. But okay, the ZK VM and the product that you're creating is an open source like library package so that any developer can take your code and then run this ZK vm. So is this, you know, in the base configuration, does this have anything to do with like Ethereum directly or Blockchain directly? Or is this. Do we start off by just creating a VM that does computer things but happens to generate proofs? **Speaker D:** Great question. So the ZK VM by itself is completely general. It doesn't have any built in features that are specific for blockchain. However, it's particularly easy to use it in a blockchain context because we do have some other stuff that we've built around it specifically to be Rails and make that nice and easy. So for example, today it's absolutely possible to get proofs out of the system and then verify them. On Sepolia, we have a smart contract that does that right now on Testnet, it's also possible to run programs inside the VM that read data from the chain and which validate that data from the chain and which therefore do so in a secure way. And so we have a number of exciting things about that that also might be fun to dig into a little bit as well. **Speaker A:** For sure, man. So I'm a little bit trying to figure out basic questions to ask you without just diving straight into the juicy stuff here. Because look, I think already how you've described it is essentially part of how I see the end game of Ethereum, which Ethereum as it exists today. The whole point is we have this prist, credibly neutral computing zone. The problem with it, and in order to maintain decentralization, which is the key product is like it's slow and terrible and expensive. **Speaker E:** Right. **Speaker A:** But in order, like I, I'm just such a big believer that like now that we've created Ethereum, we need to figure out ways to jam more like compute more tools, just more usability and like functionality into Ethereum. And so like a huge part of what I see here is like being able to project computation into the EVM via these proofs. And so one, does that, like the way I just talked about that resonate with what you guys see the purpose of Risk zero is. And two, can you just talk a little bit about why that is so important to both the company that you're building, but also the technology that we all care about? **Speaker D:** Yeah. So I think that one of the most exciting things that we're all going to get a chance to see over the next several years is exactly the scalability and performance improvements that Ethereum is sure to adopt. **Speaker E:** Right. **Speaker D:** Everyone involved in the project has got clear eyes about, you know, what needs to get done and it's really just a matter now of making the right decisions and executing them at a pace that, you know, works for people. I think that two years ago someone might have asked the question of is that going to happen? But then one year ago we saw the success of the merge, which I actually contend is one of the most impressive Internet upgrades for any protocol that we've ever seen and gives me a lot of really positive hope, I think, for the future of Ethereum and its role in the ecosystem. As far as Risk Zero and you know, the other ZK providers as well, I think that there's a number of kind of well known problems that we're all trying to solve in the best possible way and almost certainly, you know, the, the top of mind one for everyone is the scarcity of compute capacity on chain today. **Speaker E:** Right. **Speaker D:** So if you take a look at the Ethereum staking contract, you'll see that there's currently over 700,000 staked validators for Ethereum. 700,000. If we go ahead and just kind of oversimplify and say that each of those is a computer that is a phenomenally large cluster of computers. Right, this, this is, that's big money. That's, that's, that's not small beer. **Speaker E:** Right. **Speaker D:** And what we get it for that is the incredible security that Ethereum provides. There's a reason why it's the ledger of record. **Speaker E:** Right. **Speaker D:** But what we don't currently get for that is any of the scaling benefits of having such a large cluster of machines, you know, the security scales, but the compute performance currently does not. **Speaker E:** Right. **Speaker D:** And so I think that the trillion dollar next huge wave of applications kind of question here is what's the right way to solve that problem? What's interesting is a few years ago the question was is that problem solvable? So we've come a long way now. I think it's well agreed that technology like zero knowledge and more generally any kind of verifiable compute technology can solve this, will solve this. And it's now a matter of the researchers and engineers building ever better systems until finally we're at a point where we have something that makes sense for Mainnet Ethereum. So, so we're, we're in that game. You know, it's, it's a race on performance, features, compatibility, security for sure, figuring out the right design abstractions and you know, participating I think in a very large conversation that's happening right now about what that future looks like. **Speaker B:** Yeah, I think there has been and there will be like an evolution like how layer ones work I guess especially with the advent of zero knowledge proofs, they'll be more like an anchor and then it's perfectly fine if they're a bit more expensive, slower, whatever, because you can then build layer two, layer three, well, whatever on top of you just need the security of the layer one and also the flexibility to maybe get your proof system going there and make modific so your proof system. But that's going to be it. All the other stuff is going to be moved to other layers. And yeah, I think risk zero fits very well into this vision. I think we even tackle this one level lower because we're not looking exclusively into Ethereum, but we built this very general purpose machine which then you can also use for other stuff outside of crypto. And there are a lot of applications. I think people are only starting to see what applications for zero knowledge proof there are outside of crypto, in finance and insurance, whatever. There's so many, many applications for it. And yeah, they all then can be connected very nicely with crypto with zero knowledge proofs for sure. **Speaker A:** And I think like something that a lot of people who are not like Living and breathing zk, like haven't really wrapped their head around is like, we at least crypto Twitter, right, thinks about ZK as like this privacy technology. And like, I don't think any of us think of it as a privacy technology. I think we think of it as this magic that can be used for privacy but can be used for so much more. And so I think like, that's an important frame to break out of. But I'm just going to like go straight here and ask it, like, how do you guys see what you're building as like different or better than or less than a, let's say a ZK roll up. **Speaker D:** So for us, what I think we're trying to do, at least as an engineering organization, is we're not trying to race to a specific answer or specific outcome. I think that one of the things that really sets us apart is that we have a certain degree of patience if it allows us to make better engineering and scientifically informed decisions about where this technology really wants to be. It's a very abstract answer. But, you know, I think that, you know, there's some ways in which our product has already kind of changed the conversation in ZK though, and it's, it's come about in this way. So last January or February, whenever it was that we open sourced the ZK vm, we were the only people doing a general purpose instruction set architecture and really focusing on support for mainstream languages and very traditional notions of programmability. At the time, the status quo was you have kind of two options. You can build your own ZK circuits directly, which is challenging. It's fun and interesting, but it's challenging. It's not as productive as regular software engineering. It's a little bit more advanced and a little slower going. Or you could work in some of these other VMs that were designed specifically around the mathematics of zero knowledge and they needed to do that for performance or other reasons. So we came in and we had this very unusual idea at the time. We said, we actually want to make this look and feel just like a regular computer in every way except that it produces proofs about what it's done. You know, and there was a question at first, I think for that a lot of people had, which is, you know, how can you get away with that? Surely there has to be a lot of performance overhead and other kind of downsides and penalties. ZK feels very slow, so don't you need that performance? And one of this other kind of like longer horizon sort of attitudes that we took is that we're currently in a time when ZK technology itself, the mathematics and the performance is evolving super rapidly. And so we thought that if we waited long enough, the mathematics would fall into place to make those overheads inconsequential. And if we designed things right, we would be able to adopt those new developments without disrupting everything that we've already built. And so this is one of the things that this RISC V core gives us is that it's a firm separation. Everything up above that users build, it's a RISC V processor. Everything down below that we build is whatever it needs to be to achieve the best performance. So we're really trying to take that longer horizon sort of attitude. And this is sort of the opposite of, you know, building, you know, a specific ZK rollup where you have to pick a very specific model of how you think that that's going to work and be programmable. And you have to decide, are you an L1, are you an L2, are you going to be EVM based? Are you going to be something different? It's in some sense, it's a very different kind of goal. I guess I would say, yeah, let. **Speaker B:** Me maybe quickly iterate on this because that was actually, funnily enough, the point how I stumbled across RISC0 because I wanted to play around with that and maybe wanted to do a bit more than just, I don't know, doing some very artificial mathematical things that you can then bake into a circuit and then serve somehow and then risk zero. You can do that. Basically, the Prover is open source. You can write more or less any program you want as long as it's in Rust, which is currently the most supported language. You can use any libraries you want to it and then you can create a proofread. Maybe it takes too long for your use case or for your machine you're currently using, but you can play around with it and you don't have to leave the ecosystem or the tools you're familiar with. As a developer, you can just start this, use all the libraries you used before and get into it. And that's, I think, very, very interesting thing to do. **Speaker A:** Yeah, so let me try to synthesize what you guys just said with a little bit of my own flavor. But I think like, it's. So if you are at like that cusp of like really understanding what we're doing here, like you, you start to understand Ethereum as like a Turing complete computing platform that's just slow. And then like, okay, then we we're moving into this roll up paradigm where like we have the same Turing complete platform so you can do whatever computation you want. But like because we settled down to Ethereum, we can like make all these changes here. And so like essentially we have a faster and, and basically like our current paradigm is we start with slow Ethereum and as we move further out we can increase performance. And I think what you guys are saying is like that is true or not, I don't know, we can talk about it. But the problem is that you're always starting from this first principle of starting in the blockchain VM and starting with blockchain thoughts with blockchain. Whereas what Risk zero allows is like, hey, the way developers actually code is on computers, not on blockchain VMs. And so let's provide that as environment. But that also is giving us like the, the benefits of cryptography. And so maybe like the, the kind of distinction between a roll up and what you guys are building is like the, the roll up paradigm starts with like Ethereum or blockchain as first principles and builds from there. And what you guys are doing is starting with computation at first first principles and then building like zero knowledge tools to like connect it back to blockchain. Does that sound fair? **Speaker C:** It is an explanation, but it's not, maybe not a hundred percent fair. Ethereum has a very strong pull on a lot of concepts. So like Wolfgang was saying, a lot of the cryptography was there before but never really got improved until people were really interested financially in it and something like what a rollup is itself. A rollup is basically just a map produced workflow. But that term started surfacing around the time I think plasma had started to come out. And the whole idea was, yeah, let's call this a roll up because we're rolling accounts up and things like that. But essentially for Risk zero, we are doing a lot of generic things where the very fertile ground to apply those things is basically blockchain and more specifically Ethereum. And for us, the problem of verifying our proof on Ethereum actually came way after we'd established a lot of our infrastructure, a lot of our tools. So, and that was because Ethereum has a really strong pull basically. Who's most interested in all of this? It's the Ethereum crowd. Why? Because they have a long history of trying to apply zero knowledge in the context of rollups. So yeah, it's simple, it's easy to get caught up in a world where rollup is, you know, the basic foundational term. Whereas you know, a lot of these things get mixed in with the Ethereum ecosystem. **Speaker A:** Yeah, fair enough, point taken. And like the, what you guys are building is so much more applicable to crypto. But like this is the first customer and that makes sense. And I guess this conversation is just what you get when you are on a Ethereum podcast. **Speaker C:** Yeah, of course. **Speaker A:** So like let's, let's get like a little bit more concrete. Like one of the things that I was super, super excited about coming from your team is I saw on Twitter that I don't know if this is like at a production ready yet or what, but I saw someone running a version of Linux using like the Risk Zero architecture. And so can you talk a little bit about what that is, the journey to get there and then either just say it directly or we can talk about it after. But why is that one sentence Linux that is ZK provable. Such a big unlock. **Speaker D:** Yeah, so I think that there's a few folks who've actually managed to pull that off. I think that Cartesi is probably the big leader in doing ZK Linux. So you know, there's, there's a question right now that again, you know, we're the whole community is participating in, which is ultimately what are these stacks going to wind up looking like? **Speaker E:** Right. **Speaker D:** And so you know, maybe, maybe at one end we have the ZKEVM roll ups. You know, we take Ethereum as our foundation and we zkefy it and you know, this is a new platform upon which we can move some of our stuff, build some new things. Risk zero is, you know, as you mentioned, this thing that's foundationally very general. **Speaker E:** Right. **Speaker D:** But the way that normally we program Risk Zero is we write software that just runs in there directly. There's no Linux involved. We keep it really small core. And that's kind of the analogy of doing like bare metal programming, right? Very low level, you know, in principle. And then you know, maybe, maybe the right answer though is to say no, no, no, let's bring in like the entire ecosystem of open source software. And that's where you'll find, you know, the value in being able to boot up Linux. **Speaker E:** Right. **Speaker D:** That gives you everything that you can run on Linux, which is not an insignificant amount of code. So it's really exciting to see it I think because it again speaks to, you know, the scale that ZK has been able to achieve over the last couple years. In particular, it's again really far out that you can just like boot Linux and prove that that's what you did. It's not a small computation, it's not a balance transfer from one account to another. Right. Like a lot goes on to make that happen. And I think it's a real testament to this idea that we're on a path where ZK is going to be so low overhead that you won't be counting cycles, you won't be counting gas. Only if you're doing something preposterously large will will the size matter. **Speaker E:** Right. **Speaker A:** Like deploying your own risk zero instance for example. Yeah, no. And so like the way that I would describe like for when I'm trying to like tell people what Ethereum is that are not computer scientists. Like what I love to say is that like okay, the whole point of this is that the EVM is Turing complete. And what is Turing complete means? It means that mathematically kind of everything that a theory that your MacBook is capable of doing, your the EVM is capable of doing. Now it might not be easy to do it in there, it might be super expensive to do it in there. But like what Turing complete means is that mathematically we can do the same things. And like the thing that to me is just like so mind blowing about running instance of Linux with ZK provability is like my little like stupid analogy. Like is actual literal. Now like it is actually you can run anything on Linux because like when you run it on Linux it'll pop out a proof and then Ethereum can verify. Look like I didn't do the computation myself, like I don't know what kind of silicon this is running. I don't know the person who pushed the button, whatever. But I do know that based on this proof and based on like the, the input parameters that like this result came from this computation. And I think like that's huge. That's so huge that invites like there's like probably 800 solidity developers total, right? But like what this does is invite like the other like I don't know, 200 million programmers on this planet to like come participate and build solutions that leverage blockchain. **Speaker D:** That's exactly right. That's why we're so excited to have embraced Rust, a language that has a growing industrial use case around the world and which is also noted for its performance and security itself. Kind of a holy grail technology of programming languages. That's why we're also excited to unlock additional languages, you know, in the years to come. But you know, start by being good at one thing, right for sure and. **Speaker A:** Look like solidity or like the way that Ethereum works is like, we. We had solidity and then now we have like, other teams creating things like Viper or like, building adjacent to it. But like, yeah, like the actual builders, like, please focus on one thing. Like, let's let other people do interpreters and that. That kind of stuff. **Speaker B:** Yeah. But I think the computational power of Ethereum is important because I think we've seen this in the space a lot. People first start maybe with some very basic ledger that maybe this UTXO only. And they say, okay, you can move tokens, and if you dig deep, you can maybe do colored coins or some other shenanigans with utxo. But then you. I think you'll notice that you're missing stuff. You need the computational power that Ethereum has. Maybe there are some other. You don't need full theoretical Turing completeness. Maybe you could do other things. But in general that's what you need because then you have the flexibility to do those things. You don't have to worry about when developing your ZK mechanism. Oh, will that run on my SMART contract? Because essentially everything runs. Just have the freedom to do the most powerful thing there and then rely on the computational power of Ethereum, of the smart contracts and. Yeah, so I think that's definitely the way it goes. **Speaker A:** Yeah, for sure. So I've been thinking about this. I want to save kind of the end of the conversation to be about, okay, what does the world look like when you can build whatever you want on Risk zero? And what does that world? What are you enabling? But before we get there, like, let's. Let's dig a little bit deeper into, like, what it takes to project something that was computed on risk 0 into Ethereum. So, like, there's all like the. The super simple, or so not simple. There's all the super. Just generic piping of getting data back and forth. And that is what it is. But, like, can we talk about what happens on Ethereum? So let's say, like, you do something on a Risk zero instance, you are given, like, a result, which is Rex transferred 20 tokens to Tim or whatever. We don't even need to talk about tokens. But for the sake of this, let's just say that that happened on a Risk zero outside of evm, and then we're going to get the result plus a proof, and that gets sent in a transaction to Ethereum. What happens then? **Speaker D:** So actually, I'm going to turn the tables around a bit and instead of the token example, I'm going to invite Rami to talk About a different example that he explored that I think really shows the scaling and generic power of Risc 0. Maybe we can talk about the co op for a sec. **Speaker C:** Yeah, right. So the club has actually been like a dream of mine for a while because it was not actually practical to implement until I stumbled upon RISC 0. And in the club it's sort of a hybrid model or. Well, you know, we've seen this model before where you have a basically essential sequencer. That sequencer is sequencing trade orders. **Speaker E:** But. **Speaker C:** Publishing its results on chain and using a Risk zero proof to ascertain that have matched these trade orders and have moved around the balances exactly according to how a limit order book would have done those. And that is actually pretty useful because then what you, you don't run into the problems you do on a Dex, for example. And performance wise, MEV wise of course, you know, that central order book sequencer could still, you know, shuffle things around. And there are of course actually part of the club is a somewhat of an approach to tackle that. **Speaker A:** Let's put that like a little bit more into context. Right. So like the what happens if you don't have like a central limit order book that is like verifiable ftx, right? That, that is essentially like they had sorry club for those in the uninitiated central limit order book. And the idea is that in order to do trading like the sellers are on one side, the buyers on the other, everyone's sending in their orders and then somebody in the center has to coordinate that. And what our good old buddy SBF did was instead of like coordinating that honestly, he just had this like infinite balancing account where like if he needed to go negative, he would just put it in this account. And like the result of that was it all blew up. It all like fell apart because the money wasn't actually there. And Rami, it sounds like what you're saying is that like with Risk zero you can recreate FTX like all of their crazy performance and their cross collateralization and none of it was probably real. So whatever. But you can recreate like the most advanced central limit order book, run it on like you know, a super computer or right out of your, your basement or like you can have complete control of it, but because of Risk zero, the users of it can like can be assured that the clob is running honestly. Is that correct? **Speaker C:** Yeah. But to simplify a bit, the thing that we trust is basically the blockchain itself. We trust that the data we see there is valid and essentially we can see it. So what something like the Clob solution using RISC0 provides you with is basically a transaction that you can look at on chain. And that transaction would tell you that the club is actually solvent, that the money in your account in the club is exclusively allotted to you, that the club cannot take your money or use it elsewhere at all because then it wouldn't be able to provide the proof on chain and then you'd know that something is wrong. So it's not that the club has complete control, it basically only has one path to follow more or less. And that path includes no scenario where basically you lose custom or you lose your funds. So it's sort of a non custodial solution where the club's just managing the orders and doing its work, but it can't do work that harms the users. **Speaker A:** Yeah, I like so many people have like God that that's just so powerful in so many ways. Right? Because like the, the thing that is trendy to talk about is like proof of reserves and proof of liability and like I'm sure all of us like talking to some VC or some point have like off the top of our head come up with like some KZG scheme or whatever that like allows us to, to jerry rig it. But like the power of risk zero is like forget that like literally forget like building tools to like look in and make proofs and stuff. The idea is instead like we are going to have home base live in Ethereum and then yes we're going to use centralized compute to make changes to this home base. But like zero knowledge is this like tie that allows us like to take the trustlessness from Ethereum. Like you're just trusting Ethereum and I don't know, we might be just be hitting the same point over and over again. But that the central limit order book, it's huge, right? It has been the holy grail for at least defi for at least two years now. But it's like really cool to talk about what reserreal enables. So I guess we'll jump into it now. Are there any other use cases that you guys are super excited about that are kind of outside of just the simple things that we think about if you're already like an Ethereum or stuck in defi mode. **Speaker D:** So there's like a lot of exciting applications for ZK that we're really looking forward to seeing come to a broader audience. One of the first examples actually that we published was of a peer to peer video game we implemented. So there's this Classic board game called Battleship. And for folks who are familiar, right, you know, the idea is you have two players and each player kind of has their own game board and the other player can't see it. It's basically a guessing game. You take turns taking shots at pieces on the other player's board, even though you can't see. And they have to tell you whether it was a hit or not. And you know, the game proceeds accordingly. **Speaker E:** Right. **Speaker D:** So, you know, if you wanted to build that kind of thing as an Internet experience, sort of the obvious way to do it is to set up a server and all your players talk to that server and that server is keeping track of what everyone's boards look like and is answering the questions and all this kind of stuff. **Speaker E:** Right. **Speaker D:** But with zero knowledge, you can take the server out of the picture entirely and you can have two clients, you know, players talking to each other directly across the Internet and neither is able to cheat because, you know, if you ask me, hey, like, do you have anything on B4? You know, I have to produce a zero knowledge proof that my answer is faithful and relevant to the board in front of me. **Speaker E:** Right. **Speaker D:** And so this is an area that we think could be potentially very interesting for online gaming communities, especially for open source games and other kind of experiences where you have a lot of player driven content instead of centralized publisher kind of situations. But there's bringing it back though to sort of the decentralized blockchain kind of space. I think that one of the other application areas that we've gotten excited about recently is actually in the foundations of blockchains themselves. So there's another project that actually be fun to talk about. It's not out yet, but it will be out by the time all this gets posted here, which is a project that Rami Wolfgang and myself have been working on for a few weeks called Zeth Z E T H Cool. Should probably. Rami, do you want to give like a kind of a quick summary of what the idea is there? **Speaker C:** Yeah. So the Z is basically a binary built in the ZKEVM that allows you to verify the execution of an Ethereum block. So you feed it in an initial Ethereum block all of the transactions and a bunch of other authentication data. And it does what a node in the Ethereum network would exactly do to execute all of those transactions and produce a new block. And then it just gives you the block hash to tell you that after running these transactions and after processing all that input, this is the block hash that you would get. And this is very exciting for L2s and very exciting for a lot of other reasons. Because I don't know if we could say this, but we have secretly verified Ethereum main net blocks on the Ethereum testnet. **Speaker D:** Yeah, Ethereum mainnet is now an L2 onto Sepolia. It's one way to look at it. **Speaker A:** Yeah. Wait, so hold on, let me just allow my mind to explode at what you guys are talking about here. Right, so essentially, like, we have gu, right? Which is like, for the sake of argument, let's just say geth is the evm. And like, the way that it works today is that, like, you can put in a bunch of stuff and it'll build up, like the database of the EVM and then you can add new transactions and like, your guest client will spit something out. And then like, that's cool because, like, I trust my own guest client. But the way Ethereum works is then we take my block and send it to the other 700,000 computers, and every single one of those computers needs to make the exact same computations and calculations to get the same hash. Then we compare it, and then we're like, okay, good to go. What, what you guys are building with Z eth is like, my computer will get all the transactions, create the new block, and then instead of socializing that block to 700,000 computers and forcing them to recreate the same work, they can just accept it. They can run the proof, which is like, computationally speaking, like let's say 100th of actually doing the work, and we can move on from there. Is that, is that correct? **Speaker C:** For. For light clients? Light clients especially benefit from this because they don't need to mine on top of blocks. But if you need to mine on top of a block, you sort of have to also update your state so that you're able to update it further when you get new transactions. But for, like, clients, this is very useful because a lot of the work on light clients, especially in Bitcoin, has been very complicated in order to provide some kind of root of trust and have some short proof that tells you that some long suffix of the chain actually is related to the prefix somehow. Here it can all just be 1zk proof short, relatively short, and gives you all the benefits. **Speaker A:** So that's huge. And let's for a moment talk about what light clients mean, right? And I think there's light clients in Bitcoin world. And honestly, I'm not really in bitcoin world, so I don't care. But what light clients in Ethereum world mean are like today for God. I hope if you're listening to this podcast, you understand this already. But the way like when you use Metamask, your Metamask isn't actually touching Ethereum, right? Because getting into Ethereum is like a very computationally hard problem. And essentially what it is is you have to like verify 2 billion blocks to get to where we are today before you can even start. And the idea of a light client is like, instead of having to do each one of those verifications to get to where we are today and then say, I have access to the evm, we can directly access into the EVM through cryptography. And the long and short of what this means is we don't need to use infra or alchemy or centralized RPC services anymore. And so literally go look at Vitalik's writings. This is endgame stuff. This is 10 years out, we'll have Ethereum lite clients. But I believe what these three guys just told me right now is that as of today, at least on Sepoylia, like we have that for Ethereum. All right, I will take the silence and like the confident nodding as just like, you got it. **Speaker D:** Yeah, yeah. So it's, it's, it's been a pretty interesting project. I'll mention a little bit more about it and kind of how we got into it and, and what's so fun about it. So a few weeks ago, you know, we were thinking about, you know, what would be a really good test for the performance of the ZK vm. **Speaker E:** Right. **Speaker D:** We've been making a lot of improvements over the last year. We've added a number of features that are particularly exciting and differentiating. We have this thing called Continuations, which is a geeky name, but what it means is that you can run arbitrarily complex programs in there and prove them efficiently in parallel, which is like very different than how a lot of the other ZKP frameworks function today. **Speaker E:** Right. **Speaker D:** And so we sort of envision this world where you have a bunch of computers participating in the peer to peer nets for these blockchains, and where those computers are going to be kind of ordinary machines with an ordinary amount of memory and an ordinary amount of Internet connection and graphics capabilities and stuff, consumer grade stuff. That's what we think the, you know, is best to have either at home or in the data center. **Speaker E:** Right. **Speaker D:** So we picture these huge peer to peer networks full of a bunch of ordinary computers and there's an abundance of resources there. So let's go ahead and take these complicated proofs and split them across all of their so that we can shorten the time needed to get that proof completed and distributed and shared. **Speaker E:** Right. **Speaker D:** So we can achieve that finality and move the chain along. So we added these things like continuations, and it opened up this ability to just kind of write whatever program you want, no matter how complicated it kind of traditionally would seem. So we're talking about, okay, well, what would be some fun things to try out. And very quickly someone said, hey, can we build Ethereum blocks inside here? This was a really exciting kind of question because as anyone following the space over the last few years knows, there's an abundance of companies out there who want to do something that's Ethereum compatible but powered by zero knowledge. And there's a lot of reasons why that's hard and why Ethereum itself hasn't gone there yet. **Speaker E:** Right. **Speaker D:** A little over a year ago, Vitalik wrote a blog post kind of categorizing all the approaches everyone was taking. He talked about type 1, type 2, type 3, type 4, and that it's sort of a spectrum of design changes you can make from Ethereum to make stuff more ZK friendly is basically kind of the synopsis there. So a lot of folks have placed the bet that Ethereum is not ZK friendly. You got to make changes to it if you want an EVM type experience with zk, we wanted to see if that was really true. So we sat down and we're rust people. So we took a look at popular rust node software like Reth, for example, Wreath, however you want to say it. And we started looking at, you know, can we take this core logic that, you know, takes a set of transactions and current kind of a snapshot of the current state of the chain and builds a new block? Can we take that kind of core algorithm and run it inside the ZK vm? Is that something that's like possible today and pleasantly? We kind of made two observations. One, it wound up being a really easy project because there's already a bunch of rust crates that do all the heavy lifting. There's a fantastic crate revm that just does everything you need in order to apply transactions. It tells you what state changes to make. It's fantastic, it's super modular. And so I think that it was, it was under three, but let's call it four weeks of effort for us to get to a point where we were passing tests and validating real world blocks from mainnet. And in doing so, kind of without breaking a sweat like for relatively low effort, both in terms of writing the code and actually generating the proofs. So that was a pretty exciting result for us, especially given this time last year, the conversation around EVM compatible CK was how many hundreds of millions of dollars have you raised? That's such a hard engineering task. My God, right now I think it's kind of a light lift. **Speaker A:** Yeah. I can't even believe what you just said. That you built like a zero knowledge geth implementation in four weeks. I mean that, that's literally the white whale of like where we were this time last year is like one day we'll be able to. I mean, I'm trying to figure out, are we at stateless clients based on what you guys have done? Like, that's crazy. That is insane. **Speaker D:** I think it's a good step forward for it. And as far as, you know, the execute, you know, the, the time taken to build it, I mean, I think also it's a testament to Rami and Wolfgang being very good at what they do and really making it happen at just an amazing pace. But I think also it's a testament as well to the maturity of the Rust ecosystem and the benefit of again, that general purpose compute notion that we mentioned, you know, back at the beginning. If we weren't on RISC V, if we had done what other folks were doing and just designed our own architecture, we're nerds. It sounded fun, you know, like, who wouldn't want to do that? Then we wouldn't be able to just pull in a bunch of libraries to do all this. We'd be busy doing what everyone else is doing, basically, like implementing everything from scratch all the way up through, you know, just all the myriad things you have to do to make it happen. **Speaker B:** Yeah, but also we have to give a huge shout out to the Ethereum ecosystem because the stuff like ravm, like a high performance Rust EVM was there, we could use it. And that is clearly not there in other platforms or whatever because the ecosystem is not as broad and it's maybe just in typescript or whatever. And we specifically could rely on this. And this is what made it so easy to rely on our work on the zkvm, but also on the libraries and crates out there in the old Ethereum ecosystem system. **Speaker A:** For sure. I mean, like one of my, one of the reasons, like I've, I've just bet so much financially and not on Ethereum is because, like, I really believe that Ethereum is the final manifestation of open source. Like, I think that, that we were building towards this the whole time. And like, what makes Ethereum that final manifestation and not like, I don't know, Linux or whatever, is that like for the first time we benefit every time, like our open source code is used. Right. And like that has to do with like the burn and like so much of, of the organic development of Ethereum to where we are today. But yeah, I mean, I, I think like to hear this story, you have to recognize that like the work you guys did was in the two years of building Risk Zero to have this generic platform and then it's about like going to the store, picking all the parts off the shelf and then combining it together into a real application. And like, that's the magic of computer science. **Speaker D:** Absolutely. Certainly a great example of, you know, all the good stuff they taught us in undergrad about, you know, use libraries, split your work into functions. And here we are, we get to benefit from that ecosystem doing most of the heavy lifting. You know, there were a few things that we had to do that were, you know, maybe a little specific to doing it in zero knowledge. It's not as simple as taking an existing full node and just running it inside the vm, algorithmic things, it's almost that easy. But there are some things that we did to make it work securely in zero knowledge, to validate that the data being supplied into the VM is trustworthy, consistent with previous state route from the parent block, etc. But it's really very close to that. And so this is why we've taken the project and it's now out there. It's an open source thing. We want other folks to be able to take this base and run with it for their own projects and we're going to continue to build on it as well. Recently I think it came out that RISC 0, together with 01 Wabs was awarded an RFP award from Optimism to help them explore design options for their stack in zk. And this tool that we've built here is a perfect foundation for that kind of project because Optimism is of course highly EVM compatible, as are a lot of other scaling solutions out there. And so we think that this is going to be a real accelerator for the adoption of zk. And like I said, it's just out there, it's open source. Anyone can fork it, edit it, add their own opcodes, go to town. **Speaker A:** Yeah, I mean, what an amazing place to end this conversation in that like we, we have gotten to the point where ZK cryptography and like the tools being built on ZK Cryptography are so advanced and actually useful that like, Optimism, the company name for like the technology that like we thought we were going to have to use because ZK was so far out, is already like, crap. Like we need to like change what we're doing because it's here and it's ready. And so I think, like, I know at least one of you has a hard stop now, so I will hold, I want to ask you, like, okay, like, what does this mean for the. What is the purpose of Ethereum? And like what. What are the big things that we need to achieve to get there? And on and on and on and on. But for the sake of, I don't know, all of our personal lives, I will wrap it up here. So again you guys, thank you so much for your time. Of course. But like, really what you guys are building because I think without like teams like Risk zero and like looking, how do we make Ethereum a useful thing for, for modern computation? Like really what we're stuck with is NFTs and Ponzi games, right? And I think that like I have, I have nothing against somebody who wants to deploy like the next defi thing and like maybe Goldman Sachs will like want to be using Curve for trading or whatever, right? But like what I really believe is like Ethereum is not here to just like replace all the things that we hate. Like, Ethereum is here to like, change and like to, to allow new things that have never been done before. And I think for me that's been very clear for a long time. But like, what you guys are actually showing to, to the general world, like exactly how it's going to happen and so couldn't be more excited, couldn't be more proud and couldn't have any more desire to be involved. So thank you guys and before I let you go, anything you want to shout out Socials? I know you just dropped this new project, so anything, any place where we can find information on that? **Speaker D:** Yeah, so I'll start with my socials. Of course. You can find me on Twitter at intoverflow as an integer overflow and you can find Zeth our open source type 0ek zk evm at the risk0github GitHub.com risk0 with the number 0. **Speaker A:** Wolfgang, where can we find you? **Speaker B:** Also on GitHub is probably the best place like a few project there. So the GitHub handle is Wallach there? Yeah, and please do check out the RISC 0 and especially the ZEFS repository. Hope to see some great PRs coming in. **Speaker C:** I am on GitHub and Twitter as hashcach year, so just write hashcash and then ier and. Yeah. **Speaker A:** All right, again, thank you guys so much. And, yeah, I mean, I hope that I have you guys regularly on this podcast to, like, help us understand just what we're even doing here and, like, what the purpose of cryptography. And not only the purpose, but, like, the applications that are possible because of it. And I don't know. I mean, I think that nobody is really, like, at the spear tip of cryptography and this industry as much as you guys. So thank you so much for your time. And yeah, man, good luck. **Speaker B:** Thanks. **Speaker D:** Thank you, Rex. It's been a real pleasure. Hope to talk again soon.