Transcript 0:01 [funky music] So I'm here with Matt Medono at Org Summit in Singapore as part of Token twenty forty-nine, and we've had a very fun, uh, relationship online, a lot of group chats. 0:39 I've gotten to see you go from this, incubating this idea of Arch Network to landing a very significant funding for this, which is huge. Congrats on that. Thank you. Um, let's just get to it. Like at a high level, 0:54 what is Arch Network and what are you guys doing? So Arch is a, uh, it's, it's a whole execution layer, right? And layer's kind of a bad word these days, right? Because it can mean a lot of different things. 1:04 We've seen layer one, layer two, layer one point five, which may or may not, uh, have something to do with us. But, uh- [laughs]... 1:11 at the end of the day, we are a global state machine that is capable of doing Turing-complete smart contract execution- Yes... directly on Bitcoin. That was- And, uh- There you go... 1:20 and, and what we, what we really took into account is like what are the limitations that Bitcoin has, uh, uh, and what, what works for it? 1:28 And no, like taking a, a more pragmatic approach about what Bitcoin lacks and then addressing programmability from that lens really allowed us to come into this understanding of how to build this, this unique design that's, that's much different than all these other layer twos and roll-ups. 1:43 So- A year goes by. You mentioned layer one point five. I actually think that's a... I go back and forth on whether I love that or I hate it. Where do you stand right now on this, like, terminology? 1:52 So, like, with, with respect to Arch, and you have a lot of different role-dependent trust assumptions, I think layer one point five is a way that some people can really wrap their heads around it. Yeah. 2:00 Because a lot of people, when they hear the word multisig, it's like no, uh, L2 or, or side chain. 2:07 And at the end of the day, like, you don't really consider sending funds to Uniswap in, in like a smart contract-controlled wallet or, uh, in Raydium on a program-derived address to like bridging or, or interacting with a layer two. 2:20 You're using that, that, you know, multisig or that wallet for conditional execution, which is pretty much a precursor or a necessity for, uh, Turing-complete smart contracts, right? 2:30 So if we're talking about a swap on a dex, you have to have the ability to, within like a programmable way, send assets from one, uh, one, one user to another. 2:40 And, uh, and when, when on Bitcoin users are c- self-custoding their assets, it's very difficult to do that unless it's peer-to-peer. 2:46 And that, that's kinda why we have this user experience issue with, you know, the peer-to-peer type of transacting because if you look at all other programmable ecosystems, we have multi-party contracts. 2:57 You have a lot of composability, which allows for this liquidity to support this type of activity. Yeah. So we've seen that like users clearly wanna do- Yeah... like crypto activities, but on Bitcoin. 3:07 It's, I, you know, I, I think it's like the pitch is like all these things that you wanna do on these other crypto networks, GameFi, SocialFi, DeFi, whatever, you know, we're gonna do this taking it through Arch. 3:21 I understand that at a high level the market demands these things, um, but we're trying to figure out how to work it on Bitcoin, and you guys have taken an interesting approach. I think it was one of the first... 3:36 You guys were one of the first, I think, uh, teams to really embrace the Solana virtual machine on Bitcoin. Yep. 'Cause most everybody-- There's other people building like EVM stuff. You guys are SVM. 3:49 I'll ask you a question about just this idea. Does SVM, uh, provide like certain advantages over EVM, or what do you think of the landscape of virtual machines as an execution? 4:03 So, so first I'll say like we are not SVM compatible per se. Um, we are completely inspired by the SVM in the sense that we store states independently inside of UTXOs, so there's independent states. 4:15 So I m- so maybe I just misunderstood. So you're like almost mimicking SVM. Yeah. We, we've essentially rebuilt the SVM, uh, from the ground up. That's right. 4:24 So i- initially, uh, while we were working on that implementation, we were using a ZK, uh, VM, which basically the whole point of it was to produce validity proofs, and then the nodes just are validating proofs. 4:35 And what that allowed us to do is reduce the, the, like the, how heavy the nodes had to be to host each VM instance on every node. 4:42 So all they're doing, instead of running the VM, they're just validating these proofs, and when threshold is reached, the reader node is able to sign the unsigned transaction post to Bitcoin. 4:51 So once we finished our eBPF build, which is exactly the same stack that Solana is built on, we vacated the, the ZKVM, and now every node can run instances of the VM in a much more performant type of way. 5:02 Um, but simply put, the reason why we didn't choose to like just run a version of the SVM on, on Arch, I mean, on, on Bitcoin, I should say, is because the SVM doesn't take into consideration some of those shortcomings that Bitcoin has. 5:14 Yeah. There's very specific- It assumes that you're, that, yeah, SVM assumes you got blocks every- Exactly... so often, yes. You have also like the, the, the account model- Yeah... 5:21 through and through, and obviously on Bitcoin you have the UTXO model. 5:24 So take certain considerations with Bitcoin as is, um, which is what makes us so interoperable and we can read and write to Bitcoin, which is also fairly unique. 5:34 Um, so back to your original question, which is why is the SVM better than the EVM? 5:40 Uh, you have the ability to store states independently, which allows you to run transactions in parallel for non-overlapping states, which is a much more performant type of a, uh, of, of a execution environment as opposed to the sequential way, which is one transaction per block, um, which would be a pretty bad user experience for things like a dex and, and, uh, and so on and so forth. 6:00 Yeah. So especially when we have long block times like Bitcoin's.Which is this whole side conversation, like we're on a panel later today titled Slow Fi- Mm-hmm... on Bitcoin. It's, I chuckle at the term. 6:12 It's, it's funny and it- we're just- we're embracing the reality that Bitcoin block times are slow. 6:17 But, um, um, uh, jumping maybe to a question I was gonna ask towards the end, like, I was reading in the light paper that you guys published recently that on your roadmap is, um, a plan to, uh, figure out ways to increase, like, execution speed- Scalability and speed... 6:38 scalability. For sure. Yeah. So, like, inside our VM, our execution times are actually on par with Solana, so very, very fast. 6:44 Uh, the limitation or the, the bottleneck is really due to Bitcoin and their, and their, uh, block size and block times. Yeah. 6:50 So we, like, at first, we launched a, a transaction type that is settling every single state change in transition to Bitcoin every single time it executes. 7:01 So there's no rolling up of that data or compressing of that data. So I almost like, I was trying to concept, it's like, it's like a reverse rollup. It's just- Right... it's just a state publication. Exactly. Yeah. 7:11 So we're just posting a hash of the state change directly on Bitcoin, and now that is the data availability of that, you know, programmability essentially, right? Yeah. 7:20 So of course, during the process of execution, there's some, some, some elements of trust that you put on the verifier network. But once that threshold is reached, which we'll talk about- Yeah... 7:28 the threshold signature scheme in a minute, uh, why it's, it's, you know, probably the best implementation of TSS out there today, is, uh, once that threshold is reached, the nodes are able to post that data back to Bitcoin, and you're not trusting the Arch network for very long. 7:41 So that's the first transaction type, which is as close to Bitcoin as you can get. 7:46 Uh, of course, for certain types of use cases and applications, you're gonna want more performance, like perp DEXs and, uh, and, and, and also just, like, launchpads, token standards, things like that, where maybe you don't want to be anchored to Bitcoin as a developer. 8:00 You want that flexibility to choose when you wanna post that data back. 8:03 So that is, that ultimately will be managing state directly on the Arch network using the, uh, the, the Arch nodes to, to, to essentially manage that. And, uh- And that would be more like a rollup, which- Correct... 8:15 occasionally posts state to Bitcoin. Correct. But that- Or never, depending on the- Okay... on the choice of the- Okay... of the developer. 8:21 That's, that's the key, is, like, offering the flexibility for those developers to choose, "How do I want to settle? How do I want to manage the state?" 8:29 And that's very different from everyone else, where they're basically imposing that, uh, you know, that, that framework on the, the developer. 8:36 So if it's a rollup and you have to post all that data back to, to the layer one, sometimes that can be very costly, whereas sometimes they don't wanna do it at all. 8:43 They just wanna say, "Oh, I'm fine with doing this token launchpad on Arch forever." 8:48 And then you can always choose to, like, exit whenever you want, which we have some interesting things going on with the BRC20 guys, um, with their programmability module that's coming down the pipe. Yeah. 8:58 Shout out, I, I say Binary and Domo are... Yeah, this is very interesting. So it's, it's both a rollup and not a rollup- Exactly... at the same time. Yeah. 9:07 We're a, we're a global state machine that offers total flexibility to the builder. That's- Yeah. Like, they can choose the user experience that they wanna give their users, and I think that's fairly powerful. 9:17 So you talk about threshold signature scheme. I think this is actually probably one of the... I think this is probably one of the differentiators about Arch to other things. I don't think most people appreciate this. 9:27 Maybe in your experience, people, you see that people do. But I don't know, like, it's, it can be very arcane. It'd be very, like, abstract what these signature schemes actually mean. 9:38 I, myself kind of struggle to follow it, but you use Frost, Roast. Now, um, 9:47 h- just explain to me in layman's terms, for the audience who, like, maybe they're a Dejan and hasn't followed any of the conversation yet, what the hell is Frost, first of all? 9:56 So first of all, I'll say that these are highly advanced, like, tech- technology concepts that- Just, just like- No, I, I can explain it, but for the, for the viewers out there that say, like, "This is flying way over my head," uh, don't worry about it. 10:08 Like, it's, it's, uh, there are, there is only a handful of people that even know that the, the repo for Roast even is public, you know? So- See, I didn't know the repo was, was for Roast was public. 10:17 I was ac- it was funny because, like, I was talking maybe, like, six to nine months ago and I, and someone was like, "There's maybe, like, six Frost engineers out there." Right. That's right. 10:25 And there might be more now, but- Yeah. Okay, so back to Frost. Yeah, so Frost was the first, uh... I mean, like, it, there's been incremental improvements to threshold signing, right? Yeah. 10:34 You have, like, scripted multisigs, uh, and that, well, which is in threshold. Then you have scriptless, Mussig. 10:40 Uh, and, and then came Frost, which is this threshold signing, which basically means, like, you need majority consensus to ultimately sign that transaction or allow the, the multisig to do that conditional execution that's really required for, uh, for programmability or Turing-complete smart contracts. 10:56 So the, the issue with Frost is that there can be issues with live failure, meaning if nodes for whatever drop off or they don't sign, then it's, the, the chain can quite literally halt because they're unable to reach majority, and you need to have majority in order for Frost to do anything. 11:12 The, the liveness failure dilemma. Exactly. So there are a bunch of really quality projects like Botanix, like Stacks who use Frost, and that's, you know, up until recently, probably the gold standard of TSS. 11:23 The, like we mentioned, though, with the issues with live failure, we don't want the chain to be halted for any reason. 11:29 We need the chain to be s- live and active, uh, and we need nodes to be responsive in order to keep that dynamic going. The, the, the, where, where Roast ultimately comes in is it adds robustness. 11:40 It's more of a messaging protocol that gets added to, uh, Frost so that it creates these parallel signing sessions so that in the event that certain nodes go offline, there's other parallel sessions that happen so that as long as, uh, 51% does not act maliciously, you're able to reach consensus through, through these parallel signing sessions. 12:02 Okay. Let's, let's touch on that again. Sure. You have multiple sessions-And that's where you have alternate signers to be able to sign for that state to be backed on Bitcoin. Is that what Roast is or...? Yeah. 12:16 So Roast basically creates the multiple sessions, and then when a, uh, when a threshold signing session, it, you, you have like a random number of nodes that are selected to participate in that, in that signing session. 12:27 So if in the event that some of those nodes don't participate, that one session will fail. In the, the... 12:34 When you have Roast on top of Frost, you basically have the ability to create multiple signing sessions so that you have different makeups of nodes, so that if one signing session technically fails because you have nodes who are offline, another session still has the ability with other mix of, mix of nodes to, uh, to reach that consensus. 12:52 And so that would be... So that's an optimistic, uh, optimistic signing approach, I guess or- Mm. Do we...? I don't, I don't, I don't know actually if that would be an optimistic approach. 13:03 I have to, I have to check with- Yeah. I'm not-... the brain trust on that one. But, uh- Yeah... but basically, like, the way that I interpret it, and this is super layman's, and I'm not the technical guy is- Yeah... 13:11 like you have these like parallel realities, right? Yeah. One reality signing session- Yeah... 13:15 maybe it doesn't have what it needs to, to reach majority, but it's certainly not malicious because it's not 51% attack against that, that, uh, you know, malicious behavior in the signing, uh, network. 13:26 So that's like basically just one that doesn't work. That signing session is no longer relevant. 13:31 But then you also have concurrently all these other parallel signing sessions that as long as you have 51% of the nodes that say, "Yes, this happened correctly," then it's not malicious, so we can actually keep the, the, the chain going. 13:44 I think I understand. I'm gonna have to take a look at this pub- at this repo. 13:47 I mean, you, you really are correct when you say that there's so few people out there, even myself included, who really even look at Frost itself. Mm-hmm. 13:56 So then to look at Roast, which is, uh, you know, expansion of this- Right... idea is kind of a, uh, a trip. Yeah. 14:06 A lot of people don't, um, give Amin and the t- and the, and the product team credit that they deserve, I think, because, you know, Amin is extremely, extremely like avid about research. 14:15 He spends half his time coding, half his time researching. He studies bit proposals for fun, which I don't know. [laughs] Uh, that's not my idea of- I offload that to my friends who then sum that up to me. Yeah. 14:24 Yeah, ChatGPT to explain- Yeah... it back to me or something. But he's, uh, he's, he's one of the good ones. And, um, and certainly people will start to, to, to come to this realization. 14:33 I think a lot of the researchers that have done some auditing of our, of our approach and design are starting to get the picture on that. 14:39 But, uh, yeah, there's gonna be a lot of new things that, that, that we're releasing that people are gonna be like, "Holy shit," you know, "I didn't even think about that." 14:47 Okay, so this, uh, so this signing scheme, the validator network signs, uh, you, you pick who to sign, you need a threshold signature. Um, some people would kind of call this a bridge. 15:06 Why is it not a bridge? So, like, we recognize that the single points of friction and the most drop off in any of these L2 ecosystems are the bridge, right? 15:17 That's the, the, probably the single point of, like, biggest, you know, exploit that you can, that you can have. 15:23 Obviously, there's several, uh, you know, instances where we could call in the last few years where bridges have been exploited. 15:30 And at the end of the day, like, you're trusting in a completely different third party with your funds. 15:35 So you could have the best signer network in the world, like if, you know, Stacks has Frost, and if they implement Roast, that'll be probably the gold standard. 15:43 But at the end of the day, you still have to trust Xlink, or you have to trust Asigna with your Bitcoin to then go mint sBTC on Stacks. 15:50 So step one in the user journey is go to xlink.com, the bridge, and then you interact with their- Send them your Bitcoin. Yeah. Yeah. You, you interact with them. 15:59 Then you have the ability to go play around with programmability and call smart contracts on Stacks or another L2. 16:05 So Arch has basically abstracted this entire step away, where you have the ability to send funds to the smart contract directly. 16:13 And of course, the trust assumptions that you're engaging in are very, very similar, if not the same, as when you play around with the bridge, but you're still trusting the Arch network. 16:21 You're not trusting Xlink, Asigna, Wormhole, or anything like that. I think that's a really good way to put it, and actually this is how I would have defended that myself where like you... In fact, you don't almo- 16:32 You're just-- You're, you're not trusting them with your assets. You're trusting them to execute your- Correct... transaction. A similar trust assumption, but you're not actually sending them your Bitcoin yet. Correct. 16:45 It'll... Like, depending on what you're trying to do in Arch- Yeah... like, if you're like an LP in a dex, of course you have to make trust assumptions- Yeah... on the multisig and the verifier network. 16:53 So you are trusting them with your assets. However, you still don't have to s-go to a bridge and then bridge to the Arch network, and then you can send funds to the multisig. It's all in one step. 17:03 So we've essentially abstracted the bridge away and certainly don't ask users to participate with a third party. It's still Arch network and the Arch nodes that are operating on proof of stake with all of our validators. 17:15 So do you wanna talk about MEV at all? I don't know if Amin wants me to talk about it because we've got some pretty interesting things cooked up- Yeah... for MEV. 17:27 Um- Just at a high level, like, binary question, do you think MEV i-- Do you think Arch network will pr- will create opportunities for MEV? Absolutely. Yeah. 17:38 When there's execution that's happening on chain, which we certainly want to support, there's gonna be demand for a transaction ordering, which is where the MEV comes from. Yeah. 17:46 And if you ask any of the validators in the Solana ecosystem right now, their revenue from MEV is eclipsing any of the gas fee revenue that they're getting from- Yeah. 17:53 It's like ninety percent of revenue for validators is from MEV or private order flow or... Yeah. Exactly. So I think it would be very silly to tell the miners to not optimize for MEV. 18:02 Um, so we do, we will have a solution that, that addresses that for them pretty soon. Ah. So, so, so you're saying you're, you... 18:09 when you launch, you're probably going to pair this withMEV optionality or MEV tooling- Yeah... or- Optimize MEV clients, basically. Yeah. Okay. That's interesting. That's a little bit of alpha there, so. 18:20 Ooh, that is interesting. Yeah. For the miners who are listening, um, first of all, why the hell are you not paying attention to this kind of stuff? 'Cause I, if, 18:28 I, I don't know, fingers on a chalkboard hearing miners barely even talk about ordinals- Right... let alone, I think, 18:35 what is probably more important and more impactful to their bottom line, which is the data on Bitcoin and these new Bitcoin Season 2-type- Yep... activities. 18:45 Um- I think one of the reasons why they're choosing to ignore the MEV thing is because everyone else is more focused on the scalable programmability approach. Yeah. 18:53 We're more focused on unlocking programmability at the base layer, which is certainly going to drive a lot of fees back to the miners, and that's going to spike the mempool a bit and create opportunities for MEV. 19:03 So I think until the miners start getting into that fr- frame of mind, um, it may be a little bit foreign. 19:10 Uh, we've seen how ordinals and how runes have sort of added more revenue opportunities for them, but the MEV is going to be even bigger once, once the on-chain activity for the Layer 1 actually maxes out the capabilities of the chain. 19:23 I found that miners typically don't listen until they see that there o- that there is revenue. Yep. Um, so they're more reactive than they are proactive on this topic. But, um, we'll see. 19:33 I mean, I for one have a bunch of ASICs myself. I would love for those to, uh, for their margin to increase. Yep. So please post more data to Bitcoin. So, uh, does, does Arch start with a testnet? 19:45 Do you go onto like the mainnet? What does this look like? So we, uh, so our testnet is actually in beta right now. 19:50 We've distributed keys actually back the week of Nashville, so we've just been doing a lot of optimizing for the developer experience over the last month. 19:56 Uh, CLIs, SDKs, so it's really, uh, make it easier for the devs to deploy. 20:01 We are not EVM compatible, so naturally a lot of the path of progress in the developer community has been towards these EVM compatible Layer 2s, because- Fair to anybody EV, yeah... 20:09 naturally it's just copy and paste the program or the smart contract into that, into that chain. 20:14 Uh, our argument is it's gonna be easy to d- deploy those programs, but not so easy to attract TVL, um, where we may have a little bit of an advantage to get some of that TVL, but a little more friction on the developer deployment side. 20:26 So really focusing on enhancing developer experience so we can onboard the next wave of blue chips into our ecosystem. 20:33 Um, and, uh, and next week or the week after, depending on the dashboard, which is also gonna be pretty intense. Like, we're not doing a Galaxy or Zilliq quest. Like, any, no more copy-paste ideas. 20:44 Like, this is all about novel and, uh, like, thoughtful ideas. So we're gonna have a gamified dashboard that is gonna be, like, on par with Call of Duty it seems. 20:54 Like, we, we went pretty, we went pretty ham on, on this, on this incentivized testnet. And we also have an insane amount of incentives for that testnet. 21:02 So that will roll out next week, at the latest probably two weeks from now. Um, so- So we're, we're recording this on September 17th. Yep. I'll probably post this on Thursday. Hopefully the 23rd is- Yeah... 21:13 is when things roll out. Uh, we've got a nice little hype video and the dashboard and a lot of things, like creati- creatively, uh, that took a lot of time. 21:21 Um, so if not that week, then first week of October we'll be ready to go. But- I'm, uh, I'm the classic guy who te- who does everything on testnet and then, like, doesn't really do anything on mainnet. 21:31 It's just, like, my, my torture MO is, like [laughs] Like, the, I always test stuff and then never actually use it. So, um, but I look forward to using Arch Network. I wish you and Ameen luck. 21:43 You guys have been around for a long time. You've co- you've had this concept and this idea for Arch for longer than many people have been in ordinals. That's right. So. Yeah, dating back to early, early last year. Yeah. 21:54 It's, it's been, you know, it's been, uh, something that Ameen has kind of stayed true to. That's been his North Star this entire time. 22:00 When we joined the Bitcoin Startup Lab, it was, it wasn't like, uh, what should we do? We knew exactly what we wanted to do. Yeah. And, uh, it's all been about supporting his vision. Yeah. 22:09 Well, Matt, thank you so much for, uh, taking your time, and I'm looking for the SlowPi panel. Otherwise, um, where, where can people see you? Shout out the socials and, uh, yeah. Yeah. 22:18 Follow us @arch_ntwrk, uh, short for network. Um, and there's links to the Discord and the whitepaper, et cetera, all right there. Um, please read the whitepaper. 22:28 We put a lot of time and effort- It's a very beautiful whitepaper, yeah. Appreciate that. Yeah. Um, we put a lot of, spent a lot of time in whitepaper hell, you know- [laughs]... for a few weeks- Yeah... 22:37 uh, making sure that that was in line. And, uh, and then please participate in the testnet because we put a lot of work into that as well. 22:43 But if not, if for no other reason, there's gonna be lots of incentives to, to farm, um, for, for, you know, for that airdrop. So that, that's coming. Cool. Thanks. 22:53 [upbeat music]