Zorp: Nockchain's ZK Proof-of-Useful-Work Consensus - Logan Allen. Ep. 611

Transcript

Back when we were thinking of knockchain in 2023, [music] we said we want to solve distribution. We want to get more people doing [music] proofs with the Nox ZKVM than anything else as fast as possible in a self-reinforcing process. [music] And so we built a Z a zero knowledge proof of work competition, a ZK proof of work in [music] nochain. The proof of work competition is to solve [music] and and build ZK proofs. We wanted to use the consensus mechanism to drive and incentivize [music] the production at industrial scale of zero knowledge proofs. The core economic center gravity of chain is around a [music] scarce value storage instrument. And so all of the revenue generation

capabilities that are going to be built over time for data availability for programmability are about increasing [music] the monetary velocity and the usefulness of that digital. So the framing that I that I've kind of used and and that I like is thatchain is programmable sound money that scales. >> Welcome to episode of the show which talks about the technologies, projects and people driving decentralization and blockchain revolution. I'm Brian Crane and today I'm speaking with Logan Allen who is the CEO of Zorp. Zorp is a company that's been working on ZK technology and they have also launched a new proofof work ZK chain called nochain very recently. So really excited to talk

with Logan today. Um just before we get started like share a few words from our sponsors this week. >> If you're looking to stake your crypto with confidence look no further than Chorus 1. More than 150,000 delegators including institutions like BitGo, Panta Capital, and Ledger Trust Cororus One with their assets. They support over 50 blockchains and are leaders in governance or networks like Cosmos, ensuring your stake is responsibly managed thanks to their advanced me research. You can also enjoy the highest staking rewards. You can stake directly from your preferred wallet, set up a white label node, restake your assets on layer or symbiotic, or use the SDK for multi-chain staking in your app. Learn

more at chorus.1 and start staking today. >> Hey guys, I want to tell you about Nosis, a collective of builders creating real tools for real people on the open internet. Nosis has been around since 2015. In fact, it started as one of Ethereum's very first projects. And today, it's grown into a whole ecosystem designed to make open finance actually work for everyday people. At the center of it all is Nosis Chain. It's a lowcost, highly decentralized layer 1 that's compatible with Ethereum and secured by over 300,000 validators. So whether you're building a DAP, experimenting with DeFi, or working on autonomous agents, Nosis Chain gives you a solid neutral foundation to build on.

But Nosis is more than just infrastructure. It's also tools that people can actually use. Like circles, for example, lets anyone issue their own digital currency through networks of trust, not banks. And then there's Metri. It's their smart contract wallet that makes it easy to access circles, manage group currencies, and even spend anywhere Visa is accepted thanks to their integration with Nosis Pay. All this is governed by Nosis Dow where anyone can propose, vote, and help guide the network. And if you want to get involved, running a validator is super easy. All you need is one GNO and some basic hardware. To learn more and start building on the open internet, head to nosis.io.

Nosis, building the open internet one block at a time. >> Cool. Well, thanks so much for coming on, Logan. Happy to be here. >> Yeah, it's been uh I think a long time coming. I mean, I've been aware of Zorp uh since the very very beginning, but uh maybe we can start here like what what was your journey like? How did you get into well crypto I guess urbit um and sort of yeah tell us a little bit about your journey? >> Yeah, absolutely. So I have a software engineering background was programming since I was 12. Went to Georgia Tech for computer science and then I started getting really into Bitcoin and the whole cryptocurrency thing in 2016 or

so. And by 2017, I just I was in class. I was thinking, what am I doing here learning learning more math? Why am I doing this instead of just going out and working in industry trying to get Bitcoin? And so I was about to start my senior year and I decided I was going to drop out and just go move to San Francisco, work in the Bay, grab a job, and get started. Just trying to accumulate basically. >> [snorts] >> And so I've spent time at Uber doing product engineering during some of their growth periods. Spent time at Snap. And then eventually I found myself actually working more directly and adjacent to crypto at Tlon which was the research

company that built Urban. And so I spent some time there, ended up as a tech lead on their product team. And there was a lot of fun things going on at Tan at that time. It's it's hard to even describe the energy, but Ton had this very ambitious uh vision of of building a new operating system and a new decentralized internet to complement blockchains and and I really once I started interacting a lot with the technology there, I I really saw something very transformative because of how minimal and how tightly defined everything was. And in a world in which we're increasingly dependent on computing and these trustless decentralized systems for really our

everyday life, for being able to send money around, for being able to store value and in a world where centralized banking and centralized information stores are increasingly compromised at an everinccreasing rate, I really find that vision of trying to get more of our data stored and decentralized and really secure encrypted ways with really minimal security assumptions to be very compelling. And so I I spent a few years there and then uh I worked on a product studio for a little while and eventually in 2022 I founded Zorp. And initially the idea with Zorp was we're just going to do a bunch of research into how we can take one of those technologies from the knock instruction set how we can

take that and make a really really performant ZKVM a zero knowledge virtual machine that uses that instruction set and we I was able to bring on some amazing math PhDs former professors and things like this to work on it with me and Then we made some really great research results. We we recently published a paper actually that's kind of the [snorts] encapsulation of all the research we've done over the past two years into this in in June. We published an e-print and then anyway in 2023 we looked at what we had built and we said how are we going

to go to market with this right and so we decided to build chain I I've been in and around crypto for quite a while and frankly you know not everyone feels the same way about this but I've I've always really preferred the no premind proof of work ethos to the more pre-min proof ofstake ethos just because I I I think of these economic protocols as having these like core economic engines and incentives that they bring to bear in addition to the actual technological features. And I I think that there's a little bit of kind of an there's been an emphasis, let's say, on on particular incentives around basically going to market with pre-min

proof ofstake protocols because they're easier to underwrite by private placement. And I think that what we've seen over the past seven years or so since that emphasis really took hold is that there's now a lot of blockchains that all have the exact same incentives and aren't really doing any innovation on the cryptoeconomics, but they all have these like different technical flavors basically. You know, there's the AI blockchain, the fast blockchain, the storage blockchain. you know, there's so many flavors all marketing to kind of the same group of developers, all trying to be the next Ethereum, but they all have the same incentives. And so, we wanted to do something different. We wanted to we

wanted to build something that that we could that we could really get behind for a very long time. And so what we decided to do with Noctchain is is build it as a fair launch proofof work project and try to use the proofof work incentives in a novel way to incentivize real world behavior that that we wanted to see. And so of course we had this ZKVM and so I imagine you're familiar with VHS Betamax. Often often when there's a new technology the one that wins is not necessarily the best one.

It's the one that gets the best distribution. Now of course we love Noxy KVM. We think it is the best. We released this wonderful paper last last month showing formal security soundness bounds and showing that we see performance about an order of magnitude more than risk 5 VMs which is rather compelling. But we wanted to solve back when we were thinking of knockchain in 2023 we said we want to solve distribution. We want to get more people doing proofs with the Nazi KVM than anything else as fast as possible in a self-reinforcing process.

And to do that, you need to actually use incentives. You can't just have these like top down grants or anything like that. And so we built a Z a zero knowledge proof of work competition, a ZK proof of work. And so inchain the proofof work competition is to solve and and build zk proofs not not to build hashes like you see in in standard uh bitcoin. And so that was where chain was born was we wanted to use the consensus mechanism to drive and incentivize the production at industrial scale of zero knowledge proofs. I'm going to pause because you asked a very simple question and I really I really got into it. But that's that's the story. That's that's

how we got here. >> Yeah. I think there's a lot of things here that I want to dive a little bit deeper into, but maybe we can start uh with urbid a little bit because I mean urbid is something I've been involved in since uh for a long time as well and uh you know we've we've done a bunch of urbid podcasts here. Maybe two questions here. How would you describe her? What inspired you about it? And and then also I'm curious what are your biggest learnings from how you've seen the urban ecosystem evolve that you wanted to sort of apply when it comes to nochain?

>> Yeah, absolutely. So, the things that appealed to me the most about Urbbit were the idea that that really this this crypto ethos of self- sovereignty and decentralization and this idea that that you should kind of be in charge of your own destiny. You should custody your own funds. You should you should be the one that's ultimately making the decision about whether the funds move or not. You don't need custodians. this this concept for me uh I I think applies to a lot more than just money.

And and I saw when I joined Urbbit, the thing that really appealed to me the most was that they had the most expansive vision of how to apply those principles across the rest of computing. And I think that I think that that ambition to to spread the crypto ethos to a to a larger set of applicable principles, let's say to to make to make it more easy to run, say your own energy grid at home, right? Like do solar and batteries, run your own energy, don't be dependent. that that kind of almost homesteading vision with computing where you know you're growing your own food, you've got some automated systems

helping you grow your own food, you're producing your own energy, you've got some automated systems doing that and you just have these protocols that run that don't make you more dependent on everyone else but make you less dependent and in making you less dependent make you more free and more agentic. That was what really appealed to me about Urbbit and I I still find that vision very appealing uh now even even now that I'm I'm not working on the project directly and to me it's it's still about how how do we how do we bring those same principles to bear regardless of whether it's in that form factor and what are the biggest things you feel like uh you want to do different from

urb sort of lessons that you see. >> Can we swear on this podcast? [laughter] >> For sure. Go ahead. >> Okay. Okay. Uh Okay. So, uh I wanted to ask. Okay. So, so needed to ship a product to users that created a that basically showed why they would need the infrastructure that that they that they were producing the vision for.

Okay, so let's let's kind of like talk about this from first principles. If you're building a growth company and you're selling a big vision, you need to be able to provide a compelling first use case for that vision. That's a step towards the vision. Okay? As as a justification for why and and and also as a as a tool to show where you can go. So when when Elon wants to make better battery technology, he tells everybody, "I'm going to give you the best car."

And then he gets to go invest in battery technology. When Elon wants to make better rockets, he tells everyone, "Mars is really cool. Wouldn't you love to live there? Wouldn't that be so cool if we had this new place that we could go colonize and live?" And then he gets then he gets to take that vision and turn it into and and turn it into action of building tangible goods, right? Of building and making rockets better, right? So you start with selling speculative vision, then you show that you're making real steps towards capturing that energy and actually turning it into something real.

And so the thing that I think Urb has done the worst job of overall is they they have an amazing speculative vision of of what can happen with computing and they've done a really poor job of showing that they can actually take real steps towards making that real. >> Yeah. Uh certainly that has been a challenge. I agree with you. Um now let's talk about knock right cuz I think no is basically um the kind of assembly language of erit and in the beginning you know I don't think zk proofs was something that was like really consideration there but my understanding is spent or curtis spent you know five

years or something in the beginning just on knock to try to make the most simple and elegant definition of a computer And uh you know in the urbanit thing space you you'd have people then print out the whole knock definition on a t-shirt and that was kind of one of the parts. Look it's so simple it fits on a t-shirt. Like how what's your tell us more about like how do you feel about no? Why are you so excited about knock? >> Yeah. So we love so much that we then we named the whole blockchain after it. We we literally named the currency of the blockchain knock. Okay, we we are we are knock maximalists over here at Zorp and

with Knockchain. And the reason for this is that Knock is a minimal executable specification of computing in which you can do any practical thing that you want to do with a computer and specify it in terms of knock instructions. It has built-in capabilities for extension instructions. So for complex arithmetic or cryptography, you can just call out to an extension instruction in the same way that your CPU calls out to its ALU, its arithmetic logic unit for fast edition. Like you're not manually calculating in the CPU, you know, every every uh arithmetic instruction. It's basically calling out to little

specialized chip subsets on on the on the chip. So, Knock Knock is built in a way that mirrors the way that CPUs are built where you have a generic flow of logic that's minimal and can do anything turn complete computer can do and then you can call out to instructions that are extensions and knock is is agnostic to how many extensions you have or what they do but only imposes one requirement that it's a pure function. So this this is a really really key component here which is you have a minimal turning turning complete computer and then you can call out to extension instructions for any complex

arithmetic etc that you need to go really fast and so this lets you lets you do hardware acceleration of any complex logic but have an extremely consistent specification for normal computing operations. And so consistency is a really really powerful tool when building systems because consistency and having a really really minimal surface area of possible things that can occur allows you to build systems that you can actually understand and in understanding them you can cover the security holes and vulnerabilities. you can hold a mental model of what's happening in your head and you can build and work a lot

more effectively. And so the the the real things about knock that are distinct that are really unique is one it's extremely minimal and the second one is that it only uses a single data structure. Okay, so these components are going to really really matter when we start talking about ZKVMs. Okay, and the extension instruction pieces too. That's that's something that is rather unique to Knock but that most ZKVMs now use as well. So, but when Knock invented this and when Knock was introduced in 2008, this idea of these in extension instructions being included was was very very unique. No one else

was doing this. So, the the key point around this is knock is really minimal. very few instructions but turn incomplete and second no is built around a single data structure the binary tree and this single data structure it turns out is well first off it's one of the most fundamental data structures in computer science many many many things in computer science are binary trees efficient databases dictionaries etc and then second because everything is one data structure and there's a really minimal set of instructions it allows ows you to build really very efficient uh ZKVM circuit for the nonzkm and so that's that was that was the

initial intuition that that I had had in in 2022 was this is really minimal binary trees are a very wellstudied data structure with strong mathematic properties and so I thought we can probably build a really efficient ZKVM around this because ZKVM s as as a concept are a way to express computation in terms of essentially middle school arithmetic. You probably remember from um from intermediate school, right? You have you have these polomials, right? F ofx is equal to x^2 + 3 or whatever. And so a ZKVM is a way to take a computation and

express it as a relationship between polomials. So the idea is that you constrain the results of what is going to be computed using the polomials and in practice what you do is you ensure that for any given pair of rows so for any given like pair of states in the in the circuit you're going to evaluate a polomial that must evaluate to zero. So this is getting a little particular but the but in practice the way it works is you've got a computation you record all the steps that you do you put them in this big table and then you apply these polomial constraints and the polinomials have to all be zero

otherwise you did the computation wrong and zk proofs give you a way to to use this these kind of little primitives and give you a really really small proof that the computation was done correctly that can be verified extremely quickly. regardless of how really big the computation was and the proof's tiny. So let's say I've got some super huge computation I want to do, I can do it, make a ZK proof of it, send it to you, and you can verify it on your phone in like 20 milliseconds, no matter how big the computation was. This is really a powerful primitive.

It's it's kind of like so hash functions let you take a fingerprint of data no matter how big the data is and compress it into a little bitty piece that lets you verify that the data is what it's supposed to be. And so ZK proofs are kind of like a hash function but for computation instead of data. So it lets you commit to the computation. What do you think are the implications of CK proofs? like what are the use cases you're most excited about and how do you think ZK proofs are going to change the world in the long term?

>> Yeah. Okay. So the long the long term ZK proofs I think are going to totally transform finance compliance, medicine, privacy and probably things like voting also. Um, and we can we could get into the kind of like speculative idea of how these things can impact, but the general heristic is anytime that you want to be able to do something privately, but have everyone else be able to verify that it was done correctly, a ZK proof is your best tool. So, we don't really want our credit scores being leaked all over the internet every time Equifax gets hacked, but they do.

And so that type of really sensitive data that's that's the type of thing where ZK proofs would be really really useful or similarly your medical records or how you voted. It would be really really nice if every time we have an election uh everyone's not all pointing fingers and saying you cheated, you cheated, da da da. It'd be really nice if we had a public transparently verifiable mathematic representation that says everyone voted once. Everyone that voted was supposed to be able to vote and we can all verify that, but we don't know who voted for what. These types of features are uniquely enabled by ZK in a really efficient way. Now, that's kind

of the like long-term societal implications. I guess there's one more which is uh we're in a we're in a world where hacking and cyber warfare is increasingly relevant. Zero day attacks are kind of the new thing in terms of in terms of warfare. Whether whether it's the actual security zero day attack where they're, you know, hacking your information or whether it's, you know, them dropping like a cargo crate full of full of drones next to your base and then the drones come out and bomb everything. These are kind these like very sudden um very like frankly sophisticated attacks that involve technology are are the cutting edge of warfare and it really is in many ways impacting uh

critical infrastructure and critical infrastructure threat modeling. So the power grid right and water treatment facilities how do we get clean water? How do we have how do we have you know good food? How do we have power? Th this is these types of questions are obviously in some ways physical concerns. You actually have to protect these things. But these systems are increasingly digital and so securing these systems and allowing for uh introspection into these systems, verifiability that everything is going correctly.

This this is becoming increasingly important and zero knowledge proofs are a great way to be able to get that verification component. So that's all the long-term stuff. In terms of the short-term stuff, zero knowledge proofs are extremely good today for making blockchain scale. Really, really, really good. And that's that's one of the things that we're using heavily for not chain. So one of the things that blockchains are kind of one of the things I'd say is unfortunate about blockchains is today the way that blockchains uh achieve verifiability like you know if we are running a unis swap smart contract on our Ethereum nodes you guys run a lot of Ethereum nodes I know if if you're running unis

swap on it you have to be running that smart contract on every single node you're running in order to actually make the next state transition. If you want to check the next block, everybody has to run the same computation and verify it. You get verifi you get verifiability through replication of execution. You just all execute the same code. That's how you know you all got to the same answer. Well, that's really inefficient for I don't I don't know how many Ethereum nodes are running today. Do you? You probably do.

>> Um don't know exactly the number off the top of my head, but it's many. Yes. >> Yes. It's a lot, right? They're all running the same computation, right? They they all have to run the exact same thing every, you know, every new block. So, you know, you've got, I don't know, however many 30,000 computers all running the exact same thing every block. That's kind of inefficient. So, the way the way we see it is and and I mean Justin Drake and these guys are all starting starting to pitch some of this stuff too and, you know, talking about how Ethereum is going to transform over time. Um, I think they've got some fiveyear vision. Well, anyway, the the future of blockchains is point blank, you're you're going to be running the actual computation on your computer and you're going to be verifying what you

did on the blockchain. You're not going to be actually doing execution on the blockchain. That doesn't make that doesn't make very much sense. What makes a lot of sense is for you to verify a proof that you did the execution on the blockchain. And so off-chain execution, onchain verification, so that no matter how much computation you did, no matter whether you were running AI models or really sophisticated high frequency trading algorithms or crazy MEV protection or super sophisticated loan credit checks or whatever it is you're doing, whether you're running a video game, you can be doing it on your computer and then the blockchain's just verifying it was done correctly and just settling.

So you mentioned that the ZKVM that you guys wrote based on knock is is much more efficient than the other ones. >> Yeah. >> Is that particularly relevant because of the cost of generating proofs or like in what's the most important consideration for efficiency when it comes to ZKVMs? >> All right. Now, yeah, that's that's a big question. So yeah, ZKVMs are really good for scalability blockchains and they're also really good for privacy. That's one of the other one of the other big use cases. So in terms of in terms of efficiency considerations for ZKVMs.

So first off you have to understand ZKVMs are really two parts. All right. So there are two parts. The first part is you can think of it as which circuit are you running right? Are you running the knock circuit? Are you running the risk 5 circuit? Are you running the cyro circuit from start from starkqu? So it's really and so the the technical term here is um and of course different different uh different ZKVMs will make different technical decisions about things but the technical term here is going to be that's generic for all of this is the interactive oracle proof.

Okay so you're going to have your circuit which is going to be modeling some interactive oracle proof for expressing some particular computation. And so the question is is so you've got your first part which is which circuit are you running and for Starks that's going to be a randomized um arithmetic intermediate representation with pre-processing which is a crazy acronym but they shorten it to wrap. So with the start you've got your wrap and that's kind of the front end to your ZKVM. So it's which circuit you're running. Then on the back end, you're going to feed that circuit into the the ZKVM back end, which is your

polomial commitment scheme. So you've probably heard about Starks, you've probably heard about snarks, you've probably seen the term trusted setup. All right? So when people are talking about that kind of thing, they're talking about the back end. They're talking about the polomial commitment scheme. And so there's really two areas of optimization. There's the front end and the back end of the ZK view and almost all the research has has gone into optimizing the back end how to commit to these polomials and that's where most of the trade-offs come in. So do you have a trusted setup?

If you do then you can get O of one verification. You can get these itty bitty proofs that have O of one verification really tiny really really efficient but you have to trust the setup was done correctly. otherwise they can prove arbitrary statements. Or you can be a transparent commitment. In other words, there's no trusted setup. You're only trusting pure math. And in that case, the most common the most commonly used thing on the market [snorts] is is uh is Fry, which which is what Starks use. And so that's that's going to be a transparent commitment scheme. You don't have to trust anybody. It's pure math. and and with those you have larger proof sizes depending on how large the computation is and how large the circuit is. So to

answer concretely uh there's trade-offs it depends but the smaller your circuit is the more efficient you're going to be able to do the computation which is going to make proving faster and is going to make this proof smaller. And so those benefits of having smaller circuits and of having better asintoics for for building them are going to matter no matter what no matter what back end you're going into. So the work that we've done into making the knock the nokm the notch circuit really really efficient is kind of timeless in a way because it's like the core cryptography

that we can then that we can um use as a module and put into whatever backend we want. So, let's say they make this, you know, super new amazing uh proof back end, you know, let's let's say Lorito, it's like a it's a funny thing from Bang Capital where they made like Lgerero, which is another one like they made it really small, so they call it Lorito. Like, let's say that's the best thing. I don't know if that's the best thing we can port our circuit right into it and we get and we get all the same efficiency speed ups that we get right now but in the new back end and so we spent all of our time working on this on the actual circuit definition.

Okay, one thing you mentioned uh as well which is uh worth diving into I think is that you know nochain is a proof of work chain uh but the work are zk proofs now of course that is a interesting idea because in the end the idea of useful proof of work has been around for a long time right like early on people are like oh you know bitcoin is very cool but all people the miners do is to create these hashes and these hashes really don't have any function except like mining bitcoin blocks. And so of course the idea was like well what if all these miners have to do some work but that work has some other external benefit and

value besides just mining blocks. So can you explain a little bit how are the zk proofs that are produced by nochain miners how how can they be useful? Yeah, that's that's a great question. So, it's really hard to design a useful proofof work puzzle, which is why we haven't seen many many uh many attempts at it really. So the the fundamental traits that you need for a proofof work puzzle to be a secure proof of work puzzle are first off you have to be able to um verify that the puzzle was completed way faster than you can do than you can make

the puzzle in the first place. Okay, the reason for this is that you want it to be hard to spam invalid puzzles. Okay. So, you need to be able to throw, you need to be able to check that the puzzle was done properly really, really fast. That's the first thing. Then the second trait is you need it to be amortization resistant. Okay, that's a more complicated phrase, but what it but what it means is is that you don't want to be able to reuse work between attempts. Each time that you do an attempt at the proof of work puzzle, you want to basically have to start over. And sometimes you can't get all the way to amortization resistant, but you want to be as amortization resistant as you can.

And so for instance, when when uh when the ASIC boost vulnerability was published in in Bitcoin, that was an amortization um exploit is that they were able to reuse some of the work that they that they were doing between between attempts. And so uh Bitmain was able to go way faster than they should have been able to relative to using the standard algorithm. So you have to be able to check you have to be able to verify the puzzle really fast. You have to be amortization resistant at least enough and that's what you need for a proof of work puzzle. So, so this is one of the reason. So, the fact that you have to

satisfy these traits in order to make a useful proof of work puzzle at all and then of course you need to try to make it useful. This is one of the reasons why it's been so difficult for people to kind of do generalizable work in proof of work. Now, fortunately for us, ZK proofs actually satisfy a lot of these traits. Okay? So, it's a lot more expensive to make a proof than it is to verify a proof. Okay? So that's one of the first things that make it viable to make a ZK proofof work protocol at all. And then the second thing is is that if you work really hard, you can constrain down the circuit of your ZKVM enough so that there's only one valid circuit, one

valid witness for any given computation. Which means that you start with a computation, you can only make one proof with it. Okay? And if you can only make one proof per computation, it means that you have to start over again if if your attempt fails at the puzzle. Okay, so that gives you that amortization resistance. All right, so ZK proofs can can be made into valid proof of work puzzles. And so then the question becomes, which is exactly what you're saying, can we make that useful for something other than just the proof of work competition?

Well, luckily the answer is yes, you can make that useful. So uh as I mentioned knock is a touring complete VM. You can compute anything with it right and so the important point here is is that because you can compute anything with it and the algorithm for verifying the proof of work puzzle is just verifying the proof. you would be able to to theoretically provide any type of verifiable work as a proof of work puzzle result. Now inchain today in the first version that we launched we tonian cheek called it dumbnet because it was kind of like a minimal uh

shippable protocol but it's it is a useless proofof work puzzle right now. It's only used to incentivize increased prover capacity and the and the global performance competition around optimizing the nonzk which I think is very useful. It's it's hardware aligned around incentivizing people to make zk proofs faster which is useful for the whole industry. Um, but in making the actual proofs useful individually, um, it's it's actually not that big of an upgrade because currently they're they're making one proof per attempt, but they're making they're making a proof of basically a fixed computation. Now, you can imagine that that it sure would be

nice if instead say they were data availability sampling proofs. like let's say they're providing some useful service of data availability sampling or let's say they're providing a proof of transaction inclusion in the blockchain. Well, luckily there's a whole uh area of research um by by a uh by a wonderful um um cryptography PhD um Akiatus researching exactly this. how can you make uh ZK proof of work useful and and how can you understand and bound the security characteristics of it? And so there's there's some wonderful papers

on this that he's published. um we we were really uh pleased to to um to collaborate with him actually on um getting getting our research paper published on the Nox CKVM last month. But he he's spent a lot of time um over over the past, you know, many years uh publishing papers on exactly this. How can you make ZK proof of work uh useful? And so he's got a proof of necessary work paper that describes how you can once you have a a ZK proof of work that's secure in an Nakamoto consensus model, how you can actually uh use it to uh provide proofs of transaction inclusion and actually use that to power

the chain itself. So you could imagine in that model what the the chain every proof that you're doing as a part of the proof of work is actually a proof that you included a transaction in a block. And so the transaction processing is the thing being proven in the proof of work. And so the idea behind proof of necessary work is that you actually scale the chain with the proofof work competition. So the more compute power that's going into the chain for securing it, it's also powering transaction processing. So of course you know how much compute power is gone into Bitcoin. That's a lot of compute. That's a lot of energy.

Imagine if you were able to take that energy and the speed that and the speed and thoroughput of your chain was proportional to the amount of energy going into the proof of work competition. So do you think the demand for because I I guess I can see different uh avenues where this uh the CK proofs could be used? I mean one of course would be to basically say like hey look in the blockchain space there are people using ZK proofs you know kind of all over the place. go them like, hey, you should use NOX CK proofs because then you can basically earn some revenues in the form of NOC tokens and uh and you know maybe they're

also more efficient and faster and stuff like that but you know especially you have a sort of you know economic interest in adopting no ZK proof so I guess that's one the other one would be more focused on ZK proofs uh to powerchain itself. Like do you feel one of those directions do both those directions exist? And are you more bullish on one versus the other? Yeah. So I I would say I'm more bullish on using the ZK proofs to actually power specific capabilities of chain itself.

So for instance, proofs of transaction inclusion so that transaction processing uh scales up with with the with the security budget uh proofs of data availability so that you can provide basically uh data availability sampling at scale through the proofof work competition for something like a temporary blob store like you'd see from Ethereum. I I'm really bullish on these use cases and I think that I think that um the ideal situation is that you end up where knock proofs have have the competition to generate knock proofs from the ZK proof of work has generated such a massive amount of prover capacity

that individual knock proofs are extremely cheap and efficient and so there there would be no reason to even pay the protocol for the knock proofs. You see, and so the the service would be actually the proofs are just powering the protocol and you're really paying for settlement. You're paying for data availability, etc. >> So, you guys launched uh nochain in May. Um, how did it launch go? >> Yeah, launch launch was crazy, man. So, let's see. Yeah, that was that was such a crazy time. So many sleepless nights.

Um, so yeah, we we launched Knockchain. We we wanted to get it out the door as fast as we could. We've been we've been trying to get it out the door for I don't know like a year. And so we finally had had tested it enough that we were like, "Look, the whole thing works. We just got to get this shipped, get it out the door. We can keep iterating on this forever if we want to, but we're just going to get it out and and and do it for real." And and so what what we intended to do, as I mentioned, is is we didn't do a premine. So we we launched it to the public and we want Knoxchain to stimulate a global performance competition around optimizing ZK proofs.

And so the way that we kicked this off was we had published in multiple multiple of our pieces, hey, you know, the the the first Bitcoin reference clients, they they weren't optimized either. People quickly came on the scene with GPUs. people people did this optimization privately and there became this big competition and almost like war around around optimizing and and doing better in the proofof word competition and so it's of course it's 2025 now right very very different from when Bitcoin launched when Bitcoin launched only a few people even knew what hash was right and now everybody knows what cryptocurrencies are everybody knows what mining is and there's actually entire like massive server farms that

all they do is they just wait for new proof of work coins to launch and then they just go mine the heck out of them and then dump everything, right? And and so we we thought to ourselves, how can we make the fairest proofof work competition we that we possibly can in 2025 when there's all this like hostile sophisticated compute ready to be deployed and just like you know be mercenary and take everything in the dump, right? what how how can we how do we do this, right? And so what we decided to do is we decided to uh launch basically we decided to open source a few weeks ahead of launch a and and we published and talked about this in Twitter spaces for for like I don't

know like a year before we launched. So we've been talking about this for a year and and we've put out there what our business model is going to be everything. So what we ended up doing is we we launched a slow reference client. So it's like a faithful implementation of all the algorithms. It's like you know if if you take this reference client and you optimize it and you write and you make the code go faster, it will mine you a bunch of knock. And so we published this a couple weeks in advance of launch. And we said, "Hey guys, start optimizing this." And then we published a a blog post and said, "Listen, like just to be super explicit, our business model, because we didn't do a premine, our business model is we are doing

acceleration on this. Like we are we are making this go faster and we're going to be mining this with a fast client from day one. So if you want to get tokens, if you want to get knocked, you you need to optimize yours too so you can be competitive." And and so then we launched on May 21 and we had this insane flood of user. I mean there was there were 10,000 nodes join the network in like 30 minutes. It was crazy. and and um and we we got we got started and what we discovered was very very quickly was there's a massive community um particularly in um Southeast Asia

of of miners that try to join you know proof of proofof work projects particularly fair launch projects have that have a lot of uh interest in them and um that our communication around our strategy um had not gotten to them either through the language barrier or because they mostly um were listening to YouTube tutorials about how to set up the node or or whatever. Like they basically weren't engaging with our material and so they had no idea that they needed to optimize the miner to be competitive.

And so we were we were kind of floored because we got like this massive burst of attention just in the week or two coming up to launch and we kind of had no idea that it was going to be the way it was. So we got out in front of it to the best of our ability. We said, "Hey guys, listen listen. If you're if you're just running the slow code, you're not going to mind any blocks. You should you should like go get some Rust guys and write some faster code so you can be competitive." and and um honestly it pissed a lot of those people off but um but the strategy worked. So we we got a bunch of really um amazing developers and a bunch of like really dedicated and like interested guys from early bit

tensor. Uh we got some people from from who kind of came over and started a company who were uh exerbit people. We we got we got these various groups of people who who kind of like came to the call to adventure, if you will, and they optimized their minor and they got competitive really fast. The first block mined by a third party miner was block 1123 and since then, you know, like right now on the network, we're only mining 30% of blocks and the other 70% are totally unaffiliated competitive miners and we're only like 40 days into the protocol. So like basically it decentralized super fast and there's these different companies that are competing on the protocol and I've

talked to a lot of them and a lot of them I guess I haven't talked to too but basically people optimize the code and we were able to use this as a strategy to get a bunch of really useful values aligned people to join and give and basically direct all of the early token rewards to people that are actually going to work for it and not people that are just kind of coming in trying to get an airdrop and then like, you know, they don't care. They're just like here to cuz they think they're going to get rich quick and they're going to jump out, you know, and so yeah, launch was crazy, man. I had no idea what to expect.

>> Yeah, it's definitely very cool how that uh ecosystem has emerged so quickly there and how you know these different companies and I know some of them as well are are involved there. So you mentioned that you know right now it's in this kind of dumb net phase the pro the the proofs of work are not useful yet. What are the next stages in the evolution of the network? >> Yeah absolutely. So I mean one of the big things is just getting getting an ETH bridge you know getting an ETH bridge set up so that so that uh you know we can actually be connected to internet capital markets and and you know get early price discovery. I mean, at at the end of the day, right, there's

kind of this like old old meta. I mean, I it feels funny saying it's the old meta, but [snorts] uh everybody in the past few years has just been doing this thing where they like try to get like super hyped up, super high private valuations, you know, the super high FTV private valuations pre-launch, then they launch and list on an exchange and then it's down forever. you know, it's like they they try to like keep liquidity low so they can like manipulate the market and do all this like shady crap and that I don't know like I don't know why everybody's doing that. It sucks. Um everybody's sick of it. Nobody wants nobody wants that. And and so um we tried to do the exact opposite basically like you know fair launch, proof of work and we want price discovery to be

happening as fast as possible because at the end of the day like you know you live and die by the incentives. you can't like cheat the incentives. You know, if your protocol sucks, you you know, no amount of like high FTV, lowflat shenanigans is going to help. And so, you know, we believe in what we're doing. We're we're aligned around chain long term. And basically, we want early we want um we want to see what the community does and we want to see uh what happens when you connect when you connect to to broader internet capital markets. And so that's one of the that's one of the first steps is just like you know getting getting connected to to the rest of to the rest of the market. And then from there we're we're going to be adding hash time locks to support atomic

swaps uh rather shortly. Uh we we're currently working on temporary blob storage so that we can have nochain start providing data availability services. Ideally what we want to do is as I mentioned we want to move towards >> [snorts] >> um off-chain execution and onchain verification in an app rollup model where applications are issuing tokens onchain and they're they are able to perform logic and and do a lot of work offchain and use blob storage on on the chain and and and use the lock scripts and composability through intents to interact with other app roll-ups. So you can have your game or your or your club

or whatever executing offchain and then and then um interacting with assets onchain. So that's that's where we're that's where we're headed. One one way to think about this is it's basically it doesn't really make sense to try to like scale your blockchain by just centralizing and having it like do more and more replicated wasteful execution. What makes the most sense is to have as much execution happening off the chain as possible, but have the chain acting as a central coordination layer for all of that off-chain execution and providing composability between all those off-chain executions.

And so we're doing that through intents and luckily we're in the UTXO note model. And so that's that's how you do intents basically is is is by having these individual notes be able to be interacted with independently and then be able to compose atomically. And so what we're moving toward is towards providing data availability to the chain. Uh starting to provide these like very basic D5 primitives like atomic swaps and and then uh moving towards programmability. >> Okay. Okay. So this is another topic I wanted to talk about. So what is it going to look like to build applications on top of Noctchain and how does it

differ from let's say the Ethereum paradigm of how you know you create likeity smart contract then people can send transactions to interact with these smart contracts like how is it going to be different for nochain? Yeah, absolutely. So, as I mentioned, Knockchain uses the note model. So, UTXOs and so what that means is that every note has a lock on it. And so, you can spend the note if you can unlock it. So, the most common way to think of this is if you sign if you sign it, then then you can spend it if your key matches, right? That's the most simple possible lock script. Now, another lock script is a time lock.

And so, that's another one. It's like you can spin it after ever however so many blocks. That's another kind of simple lock script. But the the idea behind intents is that you can build more complex and more um more semantically meaningful conditions for spending coins. So, for instance, I could say, I'm willing to spend these coins if you trade me 100 USDC for them. That's a pretty complex condition.

And if you have these like swap conditions, as an example, that would be, you know, I'm willing to swap these coins for 100 USDC. Then you can have solvers be going through all the all the notes on the chain and saying, "Wait a second. I can I can make money by unlocking by unlocking these coins and giving these guys their 100 USDC, right? Like I'll I'll make that swap." And and so the idea of course here is that you can actually use the lot scripts as the contract.

And so the way to the way to kind of understand how this relates to nochain is not chain allows assets to com to compose with each other through lock scripts. So you can have assets interact with each other through lock scripts but the execution is happening offchain and being submitted to the chain. And because nochain is a zk native chain, we expect that for all these like complicated lock scripts, instead of having to like execute these complex computations on chain, what you're going to be doing is you're going to be verifying a proof of the lock condition on chain.

So for instance, does that does that make sense? >> Yeah, it does make sense. I mean one thing I'm curious about here is in terms of the capabilities is that you know on Ethereum you of course have like you know lending markets things like unis swap you have ds you have a lot of different types of smart contract applications do you think the this approach that nochain is taking is that going to be like as powerful >> yeah it's going to be as powerful And what what we're seeing a lot of today is particularly for complex applications

uh a lot of a lot of um a lot of complex applications are actually moving onto their own custom stacks and you you and I both know that a lot of those customs custom stacks are just Cosmos. Um but but uh but look a lot of the a lot of people are moving over to app chains. When when people first started pitching app chains, app chains were not far long enough. Like basically app chains were hyped before app chains were ready. But app chains are how these large applications are going to scale period. in regardless of what chain you're talking about, whether it's Pump Up. Fun on Salana doing their own chain or or or you know, Robin Hood deciding to do their own chain, whatever, uh

specific products that are going to do really really large amounts of of transactions, large amounts of data moving through them are going to be executing on their own. Whether we call them a chain, whether we call them an app, it it doesn't really matter. They're not going to be executing in the main chain state machine. And so what we're doing with knockchain is we have knockapps. Knocks execute offchain. Okay. If you want a central limit order book, you're going to run you're going to run it as a knockapp. If you want your AMM, you're going to run it as a knock. If you want a lending protocol, you're going to run it as a knockapp. It's going to execute offchain, but it's going to have locks on assets onchain.

And so you're going to basically post proofs to the chain and those proofs can unlock and move funds around and those and these different apps are going to compose onchain. Okay, so all of the actual assets are onchain and so as apps post proofs, they're going to be interacting with each other through the chain as a central coordinator. But the chain's not doing the execution. And it's just coordinating and composing the intent matching.

>> Very cool. Yeah, I think that is a very powerful approach. You mentioned that you guys are building, I think, decentralized exchange. Are there any other products that you guys are planning on building? >> Well, uh, so far we've seen the community's been building a bunch of products. So, I'm aware of another company, Southwest Pool Supply, that I I love the name. It's so funny. But, uh, Southwest Pool Supply, they're making a they're making a mining pool on on Knock. They call it Knockpool and and they've made an Explorer. It's beautiful. You should look at it.

Knockblocks.com. I got to say, it's got amazing metrics on it, showing minor decentralization, showing the supply schedule. I mean, these guys probably did a better job than I would have done. I mean, it it's beautiful. It looks great. I we've been seeing a massive amount of of of work starting to go in from companies that's that just sprang up basically like you know we're like we're not paying these guys. These these guys these guys are just doing doing the work because they believe in the vision. They want to participate and the proofof work protocol incentivizes them to get their hands dirty and actually work to to create value. And so for now we're we're focused on building up the protocol.

We're focused on on um you know building bridging on building and we're going to be working on uh doing a decentralized exchange um probably probably starting in uh the first half of 2026 and and uh yeah we're not we're not doing we're not going to try to do like every possible product all at once or something. We want we want to like basically um what is it? We wanna um pick our shots. You know what I'm saying? >> Yeah. Yeah, you sent me a document where you talked a bit about L1 tokens and what makes them valuable and you sort of put them into two categories. One is the store of value asset. The other thing uh

revenue generating asset, you know, I guess Ethereum would be one that you know I mean it kind of maybe fits into both buckets as well, but like it has that revenue generating component too. So where do you see knock fit in in this framework? >> Yeah, absolutely. So um I I've been I've been um writing about this and thinking about this for for a while now about how to understand and and create a valuation model for L1 assets and uh placeholder conveniently uh uh scooped me a little bit. they they published something yesterday on where they actually talk about uh they don't they the the the thrust of their essay is a little bit different than than what I've been

thinking about but they but they make the same dichotomy that that I've been looking at and so uh I was like I was like damn I got to publish this then you know I got to get this out of here like this is crazy. So the the idea is that you can understand blockchain protocols as being either primarily value storage protocols or revenue generation protocols. And it's not that you can't be both. It's that often you're optimized more toward one than the other. And so the way to think about this is a value storage protocol is a digital gold. It's like Bitcoin. And it the idea is it's a neutral store of value there.

You're a credibly neutral protocol. You're not doing things like reversing hacks and giving people their money back. You're very censorship resistant. It's very difficult to to um change your social consensus. You have an immutable supply schedule. And if someone buys the asset that it's provably scarce and they know what's they know what they're getting into and so they know that they can buy it and that basically it's going to be uh a a a hard store of value. It may fluctuate and be volatile but it's going to be scarce forever period.

And so Bitcoin is is the best example of of the digital gold that we have today. And of course, we see we see a lot of the narrative around sound money resonating and and pitching this exact thing. And and so Bitcoin is kind of the pre-minent value storage protocol today. And one of the things that makes this dichotomy really useful is well, the way that you would value something like a digital gold or a sound money is just fundamentally different than the way that you would value say like Tesla stock.

Okay? like Tesla stock is valuable, right? Or I don't know like Google stock or OpenAI stock. Like these things are these things are valuable. We agree they're valuable. Everybody thinks they're valuable. People want them. But but it's not a store of value. Just because something's valuable doesn't mean it's a store of value. Okay. So this is where I kind of bring in the differentiation around a revenue generation protocol. So on crypto Twitter, we see a lot of talk about the the revenue meta and this idea that we should kind of value protocols in terms of their ability to generate revenue through protocol services. And so the idea here is that there's kind of this dichotomy of these two different centers of gravity that protocols are naturally attracted to of whether they whether

they're primarily a value storage protocol or whether they're primarily just providing services as a protocol that they're generating revenue through. So, as you as you mentioned, um, Ethereum, right, it it does do some of both. Um, it it serves as a medium of exchange and a unit of account for the L2s and for the applications that use it and and um and does does rather well uh for it in relation to the actual uh or I should say the the valuation of Ethereum is rather high as a multiple of the revenue it generates. And part of this is because of the network effects around the way that it's used as a medium of exchange in a unit of account. We can look at a kind of an alternative example of a revenue generation protocol of Celestia.

If we look at Celestia as a revenue generation protocol, well, they really don't have any value storage capability. They only really are valued in terms of their revenue generation. There's not some big network of applications that are built on top of Celestia and using TIA, their token, as a medium of exchange or or otherwise treating it as a store of value. And they basically like people basically only value Celestia in terms of the revenue that it generates.

And so you can see the actual value of Celestia and the value of Ethereum make a lot more sense when you start to understand the difference between valuing something in terms of revenue generation versus valuing it in terms of value storage capability. So as you mentioned Ethereum does have some of both. It does have some value storage but it is primarily a revenue generation through its data availability services through its smart contract execution etc. So and of course it was the first uh programmable coin and so as a result of course it was able to develop a wonderful network effect and it's widely considered the you know number two asset. So the idea of course behind this is not to say value storage or revenue

generation that that either one is good or bad. It's to it's to have a mental framework for being able to value these assets appropriately. So I'm going to pause. Does does that make sense? Yeah absolutely. >> Nice. So, in terms in terms of notchain, notchain is primarily a value storage protocol. Notchain had no pre-mine. Notchain has an immutable supply schedule. Notchain is scarce. The only way to get it, at least right now, is is through mining it and taking part in this in this hard competition. And so notchain is going to market not not as yet another general purpose application

layer where we're going to like where the whole focus of chain is on bringing developers onto the ecosystem, right? The the notchain is going to market as a store of value. Notchain is going to market as a digital gold. And so the way to understand this is not that we are against revenue generation. As I mentioned, we're building out data availability services. We want you to be able to do um programmability. But the way to understand it is that the the frame for for the core economic center of gravity of chain is around a scarce value storage instrument.

And so all of the revenue generation capabilities that are going to be built over time for data availability, for programmability are about increasing the monetary velocity and the usefulness of that digital gold. So the framing that I that I've kind of used and and that I like is thatchain is programmable sound money that scales. So with Bitcoin, it has practically no revenue generation capability whatsoever. It's only a value storage instrument. the blockchain fees from moving transactions on Bitcoin are so minuscule as as as a percentage or as as even like just any yield [snorts] that of course people hypothesize at times about the

idea that Bitcoin if it's if its price doesn't go up at a fast enough rate that eventually the the block rewards will go so low that they won't actually serve and justify securing the protocol because it doesn't have any revenue generation capability. as a protocol. And so the way to understand this, I think, is that it all the protocols that have been launched over the past seven or eight years h have have, as I mentioned, been really focused on being proof ofstake pre-mined coins that are marketing to developers saying we're a better Ethereum. They're all revenue generation protocols primarily.

And so Bitcoin is really the value storage protocol today. And I think it's silly to think that there can't be others particularly that have differentiated characteristics. And so noctane is a value storage protocol that we intend to build in revenue generation capabilities over time. >> Okay, cool. I have one uh one more question here. So quantum computing right is something that's uh you know coming at some point and you know there's some concern about well I mean it will expected to break a lot of encryption what do you think is going to be the

effect of quantum computing on ZK and on maybe nochain in particular? >> Yeah. So quantum computing is an interesting topic because it's it's one of those things where every like everybody wants to be safe against quantum computers kind of like how you want to be safe against natural disasters and earthquakes. But there there is an open question of how how you know le let's say you don't live near any center of geological activity. How how likely is it that that you're going to be struck by an earthquake? Right?

like probably pretty unlikely. So I think in a similar way quantum computing I think quantum computing is becoming more practical over time. I I think that over time it's probably going to be able to do more stuff. Some of the breakthroughs in in being able to use quantum topological techniques to get um more more stable configurations of cubits are are rather interesting. But we're pretty far from being able to implement Shor's algorithm and actually get a sped up um you know practical implementation of of any of these attacks. That being said, of course, it takes time to upgrade protocols and it makes sense to plan in advance just in case, right? So Starks

uh depending on the the hash function that you're using for your random oracle are already plausibly postquantum secure and and uh so algebraic hashes um vary in in their in their security against these types of attacks and algebraic hashes are the ones commonly used for for securing and and u making ZKVMs go fast. So the reason is is that because they're algebraic, you can model the relationships between the hash functions more easily in terms of polomials etc.

So starts are plausibly postquantum secure depending on the hash function you use. Um and of course you can switch out the hash function reasonably easily for something like Blake 3 if you really need to. So there's a pretty easy path to taking to taking chain to be postquantum secure because we have built on starks. Uh the the main piece to think about would be the signature scheme. So signatures and and um making making sure that we have an upgrade path to a postquantum signature scheme. So, Knoxchain's not currently um secure against quantum attacks if

something just popped on the market, but we'd be able to upgrade. We we're not using any so if we were using a trusted setup as an example, it would be a lot harder to to be able to like it would be an open research question as per how to make it secure. But we're not. We're we're in a transparent scheme using very battle tested cryptography. And so the the path to the path to being totally postquantum would is is a lot clearer particularly on the timelines that that we would need to be thinking about which is like a decade.

>> Cool. Well, thank you so much for coming on Logan. That was super fascinating. I do think you guys have launched one of the most original unusual networks that has a lot of like, you know, radical design decisions you guys have made and and it's it's just like it's certainly one of the most novel things uh in crypto right now. So, I'm really excited about seeing the chain ecosystem evolve and uh it feels like it's off to a great start. So, thank you so much for coming on. >> Thanks for the time, Brian. I appreciate it.