1 00:00:00,120 --> 00:00:04,760 This is Epicenter episode 522 with the Asian Co founder of 2 00:00:04,760 --> 00:00:21,160 Scroll introducing the next generation 3 00:00:21,160 --> 00:00:24,360 of dydx and the next version of the dydx token. 4 00:00:24,680 --> 00:00:28,760 Welcome to the dydx chain. New token mechanics mean you can 5 00:00:28,760 --> 00:00:32,119 stake to secure the network. Staking is fully decentralized 6 00:00:32,119 --> 00:00:34,720 and controlled by dydx token holders. 7 00:00:35,280 --> 00:00:37,360 All fees are distributed to stakers. 8 00:00:37,720 --> 00:00:41,640 Earn rewards from using the dydx protocol with rewards plan for 9 00:00:41,640 --> 00:00:45,840 traders and early adopters too. New governance means you are in 10 00:00:45,840 --> 00:00:48,320 control. Trading has been democratized. 11 00:00:48,640 --> 00:00:51,280 You can vote on protocol improvements, token 12 00:00:51,280 --> 00:00:55,120 distributions and more. Bridge your Dydx to seamlessly 13 00:00:55,120 --> 00:01:00,480 transition to Dydx chain bridge now at bridge dot dydx dot trade 14 00:01:00,680 --> 00:01:03,480 and contribute to the evolution of dydx chain. 15 00:01:03,840 --> 00:01:06,040 Open source and community driven. 16 00:01:06,520 --> 00:01:09,520 Run your own validator validating is fully 17 00:01:09,520 --> 00:01:12,280 permissionless. Join us on our mission to 18 00:01:12,280 --> 00:01:15,160 democratize access to financial opportunity today. 19 00:01:20,520 --> 00:01:22,840 Welcome to Epicenter, the show where shocks about the 20 00:01:22,840 --> 00:01:25,800 technologies project and people driving decentralization and the 21 00:01:25,800 --> 00:01:28,480 blockchain revolution. I'm Federica ANZ, and today I'm 22 00:01:28,480 --> 00:01:35,400 speaking with Yeah, from Scroll AZKEVM, roll up on top of 23 00:01:35,440 --> 00:01:37,560 Ethereum. Yeah. 24 00:01:37,560 --> 00:01:39,320 Thank you so much for joining us. 25 00:01:39,800 --> 00:01:41,880 Hi, thanks for hosting. Nice to meet again. 26 00:01:44,000 --> 00:01:47,120 Cool. We first met at EDCON this year, 27 00:01:47,120 --> 00:01:51,640 so probably about six months ago, and had really good talk 28 00:01:51,640 --> 00:01:55,680 about kind of like, you know, L2 landscapes and L1 landscapes and 29 00:01:55,680 --> 00:01:57,840 so on. So we'll definitely get into 30 00:01:57,840 --> 00:02:01,440 that a bit later. But for everyone who doesn't 31 00:02:01,440 --> 00:02:03,400 know you, tell us a bit about your background. 32 00:02:04,320 --> 00:02:06,480 It's also glad to meet you and also your background is very 33 00:02:06,480 --> 00:02:08,759 impressive. I remember our long conversation 34 00:02:08,759 --> 00:02:10,800 debating around like here too and there one. 35 00:02:11,120 --> 00:02:15,040 OK sure, so hi everyone. My name is yeah, but I'm the Co 36 00:02:15,040 --> 00:02:17,560 founder of school. My background is more from 37 00:02:17,560 --> 00:02:20,320 academic. I will call ZK and crypto 38 00:02:20,320 --> 00:02:23,280 before. So I think before like three 39 00:02:23,280 --> 00:02:26,680 years before school I was working on purely on $0.00 proof 40 00:02:26,680 --> 00:02:29,120 research. I work on hard work solution for 41 00:02:29,120 --> 00:02:31,600 $0.00 proof because five years ago that's the biggest 42 00:02:31,600 --> 00:02:35,320 bottleneck for using $0.00 proof in practice, because a proof 43 00:02:35,320 --> 00:02:38,560 generation is just so slow. It takes like several minutes or 44 00:02:38,560 --> 00:02:42,040 several hours to generate proof for any computation any program. 45 00:02:42,280 --> 00:02:46,680 So my my first task is like use hardware like GPU and ASIC to 46 00:02:46,680 --> 00:02:49,680 make this poor generation faster to solve this kind of poor 47 00:02:49,680 --> 00:02:52,200 efficiency problem. And later I work more on the 48 00:02:52,200 --> 00:02:55,600 theoretical side, looking to how like how this kind of magic 49 00:02:55,600 --> 00:02:58,160 works, like how you generate proof, how you compress a large 50 00:02:58,400 --> 00:03:01,520 program into a very stink proof, and more like theoretical 51 00:03:01,520 --> 00:03:03,680 construction and later like you know I. 52 00:03:04,120 --> 00:03:06,640 Because you know the most fundamental problem of 53 00:03:06,640 --> 00:03:09,480 efficiency has been improved for order of magnitude. 54 00:03:09,680 --> 00:03:12,800 So that's why I moved to more on the kind of application side 55 00:03:13,000 --> 00:03:17,560 where how we can use this kind of magic technology to scale for 56 00:03:17,560 --> 00:03:21,320 privacy and for use cases. So my background is more on the 57 00:03:21,320 --> 00:03:24,840 kind of DK research side, like hardware acceleration for the K 58 00:03:25,120 --> 00:03:29,400 theoretical construction behind the K and DK applications escrow 59 00:03:29,400 --> 00:03:33,520 and mainly work on like. Taking research and some 60 00:03:33,560 --> 00:03:36,760 strategy stuff to bridge between non tech and tech. 61 00:03:38,000 --> 00:03:41,880 That's about me. Tell us about the hardware 62 00:03:41,880 --> 00:03:47,280 limitations that we currently face for ZK technology. 63 00:03:47,520 --> 00:03:54,120 So why does it place so big a burden on regular CPUs? 64 00:03:54,120 --> 00:03:56,640 And why have why did you resort to GPUs? 65 00:03:57,320 --> 00:04:00,520 Yeah, that's a great question and I think so. 66 00:04:01,000 --> 00:04:05,040 So just high level context. So the only proof is that where 67 00:04:05,240 --> 00:04:08,560 you can generate proof for so you have a program and but it's 68 00:04:08,640 --> 00:04:12,680 like too large to to execute for for verifier for example like 69 00:04:12,680 --> 00:04:15,960 I'm on a phone, I'm on a device where I can't execute the 70 00:04:15,960 --> 00:04:20,200 program myself so there's a prover which it would generate 71 00:04:20,200 --> 00:04:23,720 proof to prove that it execute the program correctly. 72 00:04:24,120 --> 00:04:26,440 This is a result, this is a proof and you only need to 73 00:04:26,440 --> 00:04:28,480 verify the proof very efficiently. 74 00:04:28,840 --> 00:04:32,720 So in this case. Prover is more like a powerful 75 00:04:32,720 --> 00:04:36,000 machine where it can execute a program and a generate proof to 76 00:04:36,000 --> 00:04:40,120 save verifies effort to to kind of re execute without re 77 00:04:40,120 --> 00:04:41,880 executing. You know the result is correct 78 00:04:42,400 --> 00:04:46,160 but the magic comes at some trade off where because you 79 00:04:46,160 --> 00:04:49,280 can't just magically save what's cost. 80 00:04:49,560 --> 00:04:52,960 So the cost is that the prover need to generate proof with a 81 00:04:52,960 --> 00:04:57,080 much larger cost. So imagine that the program like 82 00:04:57,280 --> 00:04:59,600 execution takes like for example one second. 83 00:05:00,080 --> 00:05:03,000 And then like the proof generation might takes hours 84 00:05:03,000 --> 00:05:06,920 time so which is like once on or even larger overhead to generate 85 00:05:06,920 --> 00:05:08,960 proof. So which means like initially 86 00:05:08,960 --> 00:05:12,560 prover only need to run like this program like very quickly 87 00:05:12,560 --> 00:05:15,880 but but generate proof is like you need to pay for once on the 88 00:05:15,880 --> 00:05:20,280 times overhead to just to save verifies kind of effort. 89 00:05:20,640 --> 00:05:23,960 And this kind of computation for genetic proof involving a lot of 90 00:05:23,960 --> 00:05:27,080 operations on elliptic curve is based on something called 91 00:05:27,080 --> 00:05:31,320 probabilistic shadow proof. And it would map into a lot of 92 00:05:31,320 --> 00:05:35,440 operations on evetic curve which involve large finite like large 93 00:05:35,440 --> 00:05:40,560 finite field operation like you know modular 256 bit large 94 00:05:40,560 --> 00:05:45,160 integer and that's very very expensive but luckily it's very 95 00:05:45,160 --> 00:05:49,080 very easy to parallel because. So that's why I like you know 96 00:05:49,080 --> 00:05:53,000 using GPU and and FPGA or async can make this become like order 97 00:05:53,000 --> 00:05:56,160 of magnitude faster. And we are the first one to kind 98 00:05:56,160 --> 00:06:00,320 of tackle this problem from a more academic like perspective 99 00:06:00,560 --> 00:06:03,280 and decompose proving into several components. 100 00:06:03,480 --> 00:06:07,640 Some components are like ellipse curve operations and some 101 00:06:07,640 --> 00:06:11,880 operations are like FFT over finite field and we can make 102 00:06:11,880 --> 00:06:15,480 both parts become secularly faster and then the end to end 103 00:06:15,480 --> 00:06:19,600 proof generation can be like order of magnitude faster so 104 00:06:19,600 --> 00:06:22,440 then like practically. You just run this proof 105 00:06:22,440 --> 00:06:26,440 generation over hardware and then like everything got really 106 00:06:26,440 --> 00:06:28,440 fast. And currently the state of art 107 00:06:28,440 --> 00:06:31,320 is that a lot of like companies are building different 108 00:06:31,360 --> 00:06:35,400 solutions, some are more ASIC driven like Seisig, Excel they 109 00:06:35,440 --> 00:06:38,240 are they are more leaning towards this ASIC solution which 110 00:06:38,240 --> 00:06:43,720 is highly customized and super fast, but a lot of like open 111 00:06:43,720 --> 00:06:47,200 source invitation like from from super rational from from a lot 112 00:06:47,200 --> 00:06:50,600 of teams. More like GPU based and that can 113 00:06:50,600 --> 00:06:54,480 also achieve a very very significant speed up like 10 114 00:06:54,480 --> 00:06:58,120 times or even 100 times compared with depending on the device. 115 00:06:59,200 --> 00:07:01,280 So the magic is in the parallelization. 116 00:07:01,720 --> 00:07:04,200 Yes, yes, exactly. Perfect. 117 00:07:04,440 --> 00:07:08,320 Cool. So you guys started 20/21 and 118 00:07:08,320 --> 00:07:12,160 when I say you guys, I say, I mean you and your Co founders, 119 00:07:12,160 --> 00:07:15,080 Sandy and Hai Chen. How did you 3 meet? 120 00:07:15,960 --> 00:07:18,840 Yeah. I think that's a very different 121 00:07:18,840 --> 00:07:21,880 story like we we we actually first meet meet online. 122 00:07:21,880 --> 00:07:24,400 So my background as I mentioned like I was working on academic, 123 00:07:24,400 --> 00:07:28,280 I was working on ZK purely into that rabbit hole and the other 124 00:07:28,280 --> 00:07:31,560 two Co founder Sandy and Hai Chen we work on totally three 125 00:07:31,560 --> 00:07:35,400 totally different separate area. Sandy is more working on the 126 00:07:35,920 --> 00:07:38,880 like business side, I like non tech site ecosystem building 127 00:07:38,880 --> 00:07:42,160 side and like high Chen is more leading the engineer effort. 128 00:07:42,520 --> 00:07:45,480 So it's very interesting that we we might actually through the 129 00:07:45,480 --> 00:07:48,640 ECM community. Because Sandy has been doing 130 00:07:48,640 --> 00:07:51,960 like before Scroll, he had been like, you know, doing his own 131 00:07:52,200 --> 00:07:55,840 her All Star hub and doing regulation stuff and doing a lot 132 00:07:55,840 --> 00:07:58,920 of investment in crypto. And he's a, she's more like a 133 00:07:58,920 --> 00:08:02,640 builder in the whole ecosystem and she has been paying 134 00:08:02,640 --> 00:08:05,440 attention to the whole, you know what's happening in the crypto 135 00:08:05,440 --> 00:08:07,200 world. He she noticed that there might 136 00:08:07,200 --> 00:08:10,720 be a huge opportunity to Layer 2 because Layer 2 will become the 137 00:08:10,960 --> 00:08:14,240 entry point for. Of users to enter ECM because 138 00:08:14,320 --> 00:08:17,520 ECM is just very very expensive for no more ordinary user to 139 00:08:17,560 --> 00:08:19,400 use. So there is a huge opportunity 140 00:08:19,400 --> 00:08:22,640 there to to scale using layer two And then like I would 141 00:08:22,640 --> 00:08:26,520 building like you know like thinking about how to improve 142 00:08:27,320 --> 00:08:32,159 proverse efficiency to kind of scale ECM year general like more 143 00:08:32,159 --> 00:08:35,400 general purpose way and then like high Che is more leading 144 00:08:35,400 --> 00:08:40,400 some engineer effort He he has the PhD in system from from from 145 00:08:40,600 --> 00:08:43,440 University of Washington. He has been like working on AI 146 00:08:43,440 --> 00:08:46,480 and beauty system but his system is very comprehensive. 147 00:08:46,480 --> 00:08:49,960 He builds system including like GPU compiler, it's very, very 148 00:08:49,960 --> 00:08:53,000 complete system and turning that into product. 149 00:08:53,280 --> 00:08:56,800 So it's more like what research which you know how proof of 150 00:08:56,840 --> 00:08:59,960 concept idea have you know research and architecture idea. 151 00:09:00,320 --> 00:09:03,360 One is like can actually turn this idea into implementation 152 00:09:03,400 --> 00:09:07,960 into product level like system and as I know like how to kind 153 00:09:07,960 --> 00:09:10,320 of you know like how you scale, how to fit this kind of 154 00:09:10,520 --> 00:09:14,440 technical. Component into the whole like 155 00:09:15,480 --> 00:09:19,920 ECM scaling like in diagram. Three of us are both in like I 156 00:09:19,920 --> 00:09:23,080 met Sandy through ECM community. We actually met through ECM 157 00:09:23,080 --> 00:09:26,080 research forum which I post some early idea for. 158 00:09:26,080 --> 00:09:29,640 Oh, here is you know how you can scale ECM and here is you know 159 00:09:29,640 --> 00:09:33,000 how you can Accelerate Prover and how you make that possible. 160 00:09:33,680 --> 00:09:36,640 And then like I met Hutchins through like also like our 161 00:09:36,640 --> 00:09:39,600 common friend is the ECM community and also like he he 162 00:09:39,600 --> 00:09:43,080 has been doing some like competitive programming and mass 163 00:09:43,080 --> 00:09:47,040 you know like competition. So that's how we connect and 164 00:09:47,120 --> 00:09:51,760 it's more like very organic and a very kind of you know like 165 00:09:51,760 --> 00:09:54,320 different not different way where there's the research 166 00:09:54,320 --> 00:09:57,560 forum, there's community discussion and that we we talk 167 00:09:57,560 --> 00:10:00,000 about this interesting idea and why it's possible now. 168 00:10:00,280 --> 00:10:02,920 And there are three of us just met online start to kind of. 169 00:10:03,400 --> 00:10:06,520 OK, so let's work on this. And then gradually it grows 170 00:10:06,520 --> 00:10:08,760 like, you know, totally uneffective. 171 00:10:08,760 --> 00:10:11,320 But yeah, that's a different story. 172 00:10:11,800 --> 00:10:14,760 That's a super cool origin story because kind of like purely 173 00:10:14,760 --> 00:10:17,600 online. I think it's maybe also a COVID 174 00:10:17,600 --> 00:10:19,680 story because it happened in 2020. 175 00:10:19,680 --> 00:10:22,800 One that's definitely a big yeah, yeah, So kind of. 176 00:10:22,840 --> 00:10:27,960 Real life events were kind of suspended for a while, but yet 177 00:10:27,960 --> 00:10:32,640 to me it's really amazing because kind of Co founding with 178 00:10:32,640 --> 00:10:35,680 someone. It's a very intimate 179 00:10:35,680 --> 00:10:37,880 relationship, right? So basically you have to work 180 00:10:37,880 --> 00:10:41,080 really well as a team and they're kind of finding that 181 00:10:41,080 --> 00:10:44,560 purely online, that is, that is such a cool story. 182 00:10:45,320 --> 00:10:47,360 How was it when you guys met for the first time? 183 00:10:49,320 --> 00:10:52,560 I think we still have some like share some common friends, but 184 00:10:52,560 --> 00:10:55,240 all the discussions happen like in the, you know, either 185 00:10:55,240 --> 00:10:57,760 research forum or like in the community discussion around like 186 00:10:57,760 --> 00:11:00,040 layer two skating and then like we feel like. 187 00:11:00,480 --> 00:11:03,200 The value is pretty aligned. We want to scale ECM and we 188 00:11:03,200 --> 00:11:07,080 observe that there's a very vibrant ECM research community 189 00:11:07,080 --> 00:11:08,800 there. Like that's why I'm attracted 190 00:11:09,040 --> 00:11:10,920 and there is a vibrant ecosystem. 191 00:11:11,280 --> 00:11:13,760 They are liking ECM and that that's why how Sandy is 192 00:11:13,760 --> 00:11:16,840 attracted and hygienicing this technology is so amazing. 193 00:11:16,840 --> 00:11:18,080 I want to turn that into a system. 194 00:11:18,080 --> 00:11:21,560 So it's like there's still some common friends, but you know, 195 00:11:21,560 --> 00:11:25,200 it's more like community thing and then like people meet and 196 00:11:25,360 --> 00:11:29,120 like got to know each other and yeah, yeah, but you guys have 197 00:11:29,120 --> 00:11:31,520 met in person, right? Yes, yes. 198 00:11:31,520 --> 00:11:35,000 It's only after I think the the earliest in person meet up is in 199 00:11:35,000 --> 00:11:37,840 devconnecting Amsterdam, where that's the first time we 200 00:11:37,880 --> 00:11:39,320 followed. That was 2022. 201 00:11:40,520 --> 00:11:44,920 Yes, yes, I think so cool. What Scroll puts front and 202 00:11:44,920 --> 00:11:46,920 centre is that it is EVM compatible. 203 00:11:47,040 --> 00:11:51,520 What drove you to this decision Because I mean we recently spoke 204 00:11:51,520 --> 00:11:57,520 with a number of other ZK roll ups and some of some of which 205 00:11:57,520 --> 00:12:02,480 kind of also went this route like Hamas with with Geordie, 206 00:12:02,480 --> 00:12:07,760 now Polygon, ZKEVACKEVM, but others haven't. 207 00:12:07,800 --> 00:12:13,720 So kind of like for instance we recently had on Aztec and Ellie 208 00:12:13,720 --> 00:12:18,520 from Starknet and they were adamant that the efficiency 209 00:12:18,520 --> 00:12:23,440 losses that you suffer from doing an op code by op code 210 00:12:23,640 --> 00:12:26,640 based translation is not worth it. 211 00:12:27,320 --> 00:12:30,800 So how? How did you kind of settle on 212 00:12:31,000 --> 00:12:34,120 this design choice? Yeah, that's a great question. 213 00:12:34,120 --> 00:12:39,000 So I think a big, a big part of that is that like you know think 214 00:12:39,000 --> 00:12:42,040 about your motivation from first principle, like you know why you 215 00:12:42,040 --> 00:12:43,640 want to even build a layer two solution. 216 00:12:43,960 --> 00:12:47,920 So I think for me that we observe that ECM is congested 217 00:12:48,080 --> 00:12:51,800 and it's very expensive and applications on ECM want to find 218 00:12:51,800 --> 00:12:54,040 a place where it it's very cheap. 219 00:12:54,400 --> 00:12:57,520 It's auto secure. It has a security from ECM So 220 00:12:57,520 --> 00:13:02,120 that's why we think you know like reducing the effort that 221 00:13:02,120 --> 00:13:06,120 applications can find such a solution either way to go and 222 00:13:06,120 --> 00:13:09,320 also like I think from and like. So that's that's our first 223 00:13:09,320 --> 00:13:13,280 motivation, like why we ought to build a very seamless scaling 224 00:13:13,280 --> 00:13:16,400 solution that all the developer, all the users can reuse 225 00:13:16,520 --> 00:13:21,080 everything they have to migrate to, to scroll in a seamless 226 00:13:21,080 --> 00:13:22,920 manner. And the users can reuse all 227 00:13:22,920 --> 00:13:26,760 their familiar like toolings and developer can use their kind of 228 00:13:26,960 --> 00:13:30,520 foundry and all those tools directly and also like. 229 00:13:31,000 --> 00:13:33,680 So that's the first intuition where we don't want anyone to 230 00:13:33,680 --> 00:13:37,600 change any of their kind of habit to only get the benefits 231 00:13:37,600 --> 00:13:43,840 which is faster, cheaper with a faster like pre confirmation 232 00:13:43,840 --> 00:13:47,480 time and high throughput. So that's our first intuition. 233 00:13:47,880 --> 00:13:50,600 And second thing is that I think it's actually from the security 234 00:13:50,600 --> 00:13:54,480 perspective where EVM has not only has a ecosystem but also 235 00:13:54,480 --> 00:13:59,240 has proven itself from years of of of time where you know like 236 00:13:59,240 --> 00:14:03,160 all the all the application deployed on this model like 237 00:14:03,160 --> 00:14:05,400 hasn't been, there hasn't been any any problem with that. 238 00:14:06,000 --> 00:14:08,720 So that's why like we think inheriting the security from 239 00:14:08,720 --> 00:14:11,200 this kind of test of time model is very important. 240 00:14:11,400 --> 00:14:14,760 Even like we we are reusing to the code level where we are 241 00:14:14,760 --> 00:14:18,760 trying to reuse a code. They kind of know the invitation 242 00:14:18,760 --> 00:14:20,720 from ECM. We are trying to use Go ECM 243 00:14:20,720 --> 00:14:25,000 which is ECM's client invitation to generate block to generate 244 00:14:25,040 --> 00:14:28,240 our block to make sure that the behavior is exactly the same. 245 00:14:28,520 --> 00:14:32,320 So using that it's definitely increases the security and also 246 00:14:32,320 --> 00:14:35,120 because developer don't need to change any layer of their code. 247 00:14:35,360 --> 00:14:39,400 So like imagine like you know if every layer to require you to 248 00:14:39,400 --> 00:14:42,240 make a significant changes to our code, then like maybe 249 00:14:42,520 --> 00:14:46,120 largely JIT application like Uniswap Avi might worry that you 250 00:14:46,120 --> 00:14:48,040 know. Why should I migrate to this 251 00:14:48,040 --> 00:14:51,720 chain with some risk in changing my code and we're doing this re 252 00:14:51,720 --> 00:14:55,440 audit so that's why I like we think this is very important and 253 00:14:55,560 --> 00:15:00,200 and also like I think EVM just have way more like toolings than 254 00:15:00,200 --> 00:15:03,840 like any other VM has like even if you think you know software 255 00:15:03,840 --> 00:15:08,120 has has put a lot of effort in kind of Carl but like if you 256 00:15:08,120 --> 00:15:10,560 think about how many invitation how many toolings you can use to 257 00:15:10,560 --> 00:15:15,080 deploy on on Carl that like. It's very limited compared with 258 00:15:15,080 --> 00:15:19,200 with the whole EVM ecosystem. So that's another thing which 259 00:15:20,000 --> 00:15:22,880 you know like enhard the ecosystem build your own network 260 00:15:22,880 --> 00:15:26,360 effect and then like only then like you can think about you 261 00:15:26,360 --> 00:15:30,520 know what things you can you can provide some actual value to to 262 00:15:30,520 --> 00:15:33,440 the developer in your community. And I think one last last thing 263 00:15:33,440 --> 00:15:38,200 is that this even distinguish us from several other like language 264 00:15:38,200 --> 00:15:41,720 compatible, they grew up is that we we are like by core level 265 00:15:41,720 --> 00:15:44,560 compatible with EVM. So which means like per OP code 266 00:15:44,560 --> 00:15:48,720 we will have some circuit mapping to kind of prove on the 267 00:15:48,720 --> 00:15:52,920 bicycle level and so and reuse all the kind of invitation like 268 00:15:52,920 --> 00:15:54,760 Go ECM to kind of generate our block. 269 00:15:55,040 --> 00:15:59,360 So this guarantee us that every time ECM is doing any upgrade, 270 00:15:59,600 --> 00:16:03,560 it's very easy to apply to our chain like for example like you 271 00:16:03,560 --> 00:16:06,600 know there is, there is some kind of Eips, there's hard fork 272 00:16:06,840 --> 00:16:10,480 and it's easy for us to adopt those changes if we are EVM, but 273 00:16:10,480 --> 00:16:12,680 if you are. Totally work on different paths. 274 00:16:12,760 --> 00:16:15,960 It's very hard to kind of follow what is what what the ECM layer 275 00:16:15,960 --> 00:16:19,040 one is doing, adopt all the innovations from the ECM 276 00:16:19,040 --> 00:16:21,440 community. So even if like you know there's 277 00:16:21,440 --> 00:16:25,360 so many innovations like new pre compilers, new discussions 278 00:16:25,360 --> 00:16:29,520 around like you know ECM layer one and we can directly take all 279 00:16:29,520 --> 00:16:31,560 those innovations and apply that to layer 2. 280 00:16:31,840 --> 00:16:35,560 Even some like infrastructure like 43C7, PBS, all those kind 281 00:16:35,560 --> 00:16:38,320 of great ideas like MEV, all those kind of infrastructure to 282 00:16:38,320 --> 00:16:40,480 solve those problem can be reused on. 283 00:16:40,720 --> 00:16:42,800 Or scroll directly. So that's why like you can 284 00:16:42,800 --> 00:16:47,000 always, you know stay up to date with being very ECM aligned, 285 00:16:47,320 --> 00:16:50,560 reuse all the innovation from the research community without 286 00:16:50,560 --> 00:16:54,880 too much fragmentation. And eventually I think you know 287 00:16:54,880 --> 00:16:59,160 later might become especially us might become the the the kind of 288 00:16:59,920 --> 00:17:02,920 like we can go you know one one step ahead of ECM to test some 289 00:17:02,920 --> 00:17:07,200 kind of Eips and adopt some innovations and then like if 290 00:17:07,200 --> 00:17:08,920 it's proven to be successful then like. 291 00:17:09,319 --> 00:17:12,319 Maybe ECM has a larger possibility to to kind of adopt 292 00:17:12,520 --> 00:17:15,280 to push this innovation. So always stay ahead, always 293 00:17:15,280 --> 00:17:19,440 stay on top of this kind of nice solutions and to benefit the 294 00:17:19,480 --> 00:17:23,079 whole ECM community. So yeah, there are multiple 295 00:17:23,079 --> 00:17:25,880 reasons but that's all like you know, I can, I can think about 296 00:17:26,000 --> 00:17:29,080 definitely like TLDI that developer and user expenses, the 297 00:17:29,080 --> 00:17:33,240 same security model, the same ecosystem and toolings are more 298 00:17:33,240 --> 00:17:35,040 vibrant. And last thing is like you know 299 00:17:35,040 --> 00:17:38,560 stay always ahead on this innovation and research. 300 00:17:39,120 --> 00:17:43,720 And benefit eventually benefit the whole ECM community and I 301 00:17:43,720 --> 00:17:49,040 hear that 100%. But can I just kind of poke into 302 00:17:49,040 --> 00:17:52,680 this a bit more? So if if you kind of look at, 303 00:17:53,840 --> 00:17:57,080 what would you say the efficiency gains you could get 304 00:17:57,080 --> 00:18:02,880 from not using a ZKEVM would be and kind of, is there any way 305 00:18:02,880 --> 00:18:07,000 kind of to make up the efficiency you're losing with 306 00:18:07,000 --> 00:18:12,000 respect to other roll ups that are using more optimized 307 00:18:12,200 --> 00:18:15,240 languages? Yeah, that's a great question. 308 00:18:15,240 --> 00:18:19,120 So I think that's actually the biggest like reason why you know 309 00:18:19,400 --> 00:18:22,440 like so many row OPS want to choose other virtual machine 310 00:18:22,440 --> 00:18:25,320 model because they used to think like you know being by code 311 00:18:25,320 --> 00:18:29,360 level compatible building such as the KVM will have a 312 00:18:29,360 --> 00:18:32,080 significant larger overhead. Maybe they they are imagine like 313 00:18:32,480 --> 00:18:35,600 10 times, maybe like 100 times, maybe like larger overhead than 314 00:18:35,600 --> 00:18:38,040 they are. They are kind of the key virtual 315 00:18:38,040 --> 00:18:41,200 machine. But the reality is that I think 316 00:18:41,240 --> 00:18:44,560 they didn't expect even like us do they expect like the prover 317 00:18:44,560 --> 00:18:48,960 technology has been proved so much I I think two years ago, 318 00:18:48,960 --> 00:18:51,000 two or three years ago. I think when we start school, 319 00:18:51,240 --> 00:18:54,720 compared with five years ago where the technology of the K 320 00:18:55,200 --> 00:18:58,640 lies, it's like the the efficiency of poor has been 321 00:18:58,640 --> 00:19:02,840 proved for three order of magnitude like by this kind of 322 00:19:02,960 --> 00:19:06,320 advanced proving system by the underlying hardware 323 00:19:06,320 --> 00:19:08,480 acceleration. By proof recursion. 324 00:19:08,720 --> 00:19:11,880 So the efficiency has already been improved for like 1 sum of 325 00:19:11,880 --> 00:19:14,400 times. So that's why I think compared 326 00:19:14,400 --> 00:19:17,240 with different ZQ virtual machine, I don't think ZQ VM by 327 00:19:17,240 --> 00:19:19,520 code level ZQ VM had that much overhead. 328 00:19:20,280 --> 00:19:23,600 I think had most probably like two or three times even like 329 00:19:23,600 --> 00:19:26,800 sometimes like can be even lower and that's only talking about 330 00:19:26,800 --> 00:19:30,560 the poor cost which is not the dominant cost for a layer 2 331 00:19:30,560 --> 00:19:32,480 transaction. So think imagine like you know 332 00:19:32,480 --> 00:19:35,600 for our kind of layer 2 transaction if you want to post 333 00:19:35,600 --> 00:19:38,840 data on chain. That might takes the majority of 334 00:19:38,840 --> 00:19:43,760 the cost which is over 90% of the the cost and then like among 335 00:19:43,760 --> 00:19:47,800 the kind of rest of 10% of time. You know like maybe using some 336 00:19:47,800 --> 00:19:51,440 other Z key VM can save the proving cost for like 2 * 3 337 00:19:51,440 --> 00:19:55,000 times and that's only portion 1 portion of this 10%. 338 00:19:55,280 --> 00:19:58,560 Which means like you know it won't differ that much even if 339 00:19:58,560 --> 00:20:00,040 you are using other Z key virtual machines. 340 00:20:00,480 --> 00:20:03,640 But the the the hugest loss is that you lose compatibility. 341 00:20:04,160 --> 00:20:06,880 And you have to rebuild your ecosystem from scratch. 342 00:20:07,240 --> 00:20:09,920 So and you will really suffer from that. 343 00:20:10,280 --> 00:20:13,760 So I think that's why I don't see kind of too much loss on the 344 00:20:13,760 --> 00:20:17,840 on the on the poor side and especially with the technology 345 00:20:17,840 --> 00:20:21,640 keep improving and we predict like the the technology will 346 00:20:21,680 --> 00:20:25,560 still improve for another 10 times like in my prediction is 347 00:20:25,560 --> 00:20:28,600 like in six months or one year we can make a 10 times even 348 00:20:28,600 --> 00:20:32,320 faster the KVM. So it's really not that painful 349 00:20:32,400 --> 00:20:34,720 as as people predict like five years ago. 350 00:20:35,400 --> 00:20:38,200 So, so you think it's kind of, is it fair to kind of compare 351 00:20:38,200 --> 00:20:46,480 this to say no longer building applications in C++ just because 352 00:20:46,480 --> 00:20:50,560 I mean just because C++ may be a little bit more efficient at 353 00:20:50,560 --> 00:20:53,600 some things and kind of the convenience and the developer 354 00:20:53,600 --> 00:20:56,960 mindshare that you kind of get with more? 355 00:20:58,560 --> 00:21:01,280 Common developer languages more than makes up for this. 356 00:21:01,800 --> 00:21:04,840 So are you talking about like using SIPS also to write smart 357 00:21:04,840 --> 00:21:07,760 contract versus Solidity or? Like, Oh no, no no, not not 358 00:21:07,760 --> 00:21:11,960 smart, not smart contract, just general general code. 359 00:21:12,000 --> 00:21:16,720 I mean, so basically I went to university a while ago, and back 360 00:21:16,720 --> 00:21:21,440 then whenever we kind of took home large batches of data, 361 00:21:21,760 --> 00:21:25,760 obviously C++ was kind of the go to thing to kind of use, just 362 00:21:25,760 --> 00:21:27,800 because it was much more efficient than Python for 363 00:21:27,800 --> 00:21:31,560 instance. But you you're saying that kind 364 00:21:31,560 --> 00:21:36,240 of using Python And other more convenient languages for 365 00:21:36,240 --> 00:21:40,040 developers, this is kind of the parallel, right? 366 00:21:40,040 --> 00:21:43,480 Kind of it's kind of like you you, you you're saying kind of 367 00:21:43,640 --> 00:21:48,120 the technology has made such tremendous advancements that 368 00:21:48,120 --> 00:21:51,200 kind of the fact that two or three that you're possibly 369 00:21:51,200 --> 00:21:52,880 losing doesn't really matter much. 370 00:21:54,400 --> 00:21:56,800 Yeah, in some sense, yes. Like we want to keep the kind of 371 00:21:56,800 --> 00:22:01,160 Python level developer experience but magically reduce 372 00:22:01,160 --> 00:22:04,560 the overhead of back end to some degree that you know it doesn't 373 00:22:04,560 --> 00:22:07,240 matter like you know you optimize safely for C plus 374 00:22:07,240 --> 00:22:10,080 because that doesn't really influence your overall cost in 375 00:22:10,080 --> 00:22:14,200 some sense yes but it's more similar to like I think other VM 376 00:22:14,200 --> 00:22:19,440 are trying to kind of like compel like I, I I I think like 377 00:22:20,120 --> 00:22:24,000 it's very similar but one thing which is quite different is that 378 00:22:24,480 --> 00:22:28,840 if you run C++ and Python like the the the complexity of the 379 00:22:28,840 --> 00:22:33,000 underlying like CPU is like still like influence on 380 00:22:33,000 --> 00:22:35,160 efficiency. But here like especially if 381 00:22:35,160 --> 00:22:37,440 you're looking at the transaction cost, it's not just 382 00:22:37,440 --> 00:22:40,960 execution cost, it's also like data cost, which means like make 383 00:22:40,960 --> 00:22:43,880 this kind of difference become even smaller because it's like 384 00:22:44,080 --> 00:22:48,240 5% or more all the cost and then you save this 5% to some degree. 385 00:22:48,640 --> 00:22:52,440 But you know you you give up like this, compatibility either 386 00:22:52,440 --> 00:22:56,000 user, but the transaction cost overall will be similar, yeah. 387 00:22:56,560 --> 00:22:59,280 Yeah, perfect. I want to talk about kind of how 388 00:22:59,400 --> 00:23:02,960 the cost for an L2 transaction is made-up later and kind of 389 00:23:02,960 --> 00:23:04,600 talk about data availability then. 390 00:23:04,600 --> 00:23:08,480 So kind of I I would like to table this for now, let's talk 391 00:23:08,480 --> 00:23:13,760 about kind of the the creation of scrolling a little bit more. 392 00:23:13,760 --> 00:23:17,080 So I remember the ECC presentation where where Jody 393 00:23:17,280 --> 00:23:22,680 for the first time. Talked about building a OP code 394 00:23:22,680 --> 00:23:27,120 by OP code transposition into a ZKEVM that was in June 2021. 395 00:23:27,160 --> 00:23:30,000 Blew everyone's minds. How did you? 396 00:23:30,320 --> 00:23:32,120 How did you guys fit into your timeline? 397 00:23:32,120 --> 00:23:35,800 Because I know scrollers. Scrollers also founded in 2021. 398 00:23:36,640 --> 00:23:39,080 Yeah that's a that's a really great question. 399 00:23:39,160 --> 00:23:41,360 So I think a lot of people are like maybe don't know the the 400 00:23:41,400 --> 00:23:43,880 the story behind. I think, I think while we 401 00:23:43,880 --> 00:23:48,800 started, we start in early 2022 in January, I think that's the 402 00:23:48,840 --> 00:23:51,520 earliest we have this idea for we want to build a layer 2. 403 00:23:51,920 --> 00:23:54,840 But the first idea of score is that because we have this 404 00:23:54,840 --> 00:23:57,960 hardware idea in mind, so we want to design some 405 00:23:57,960 --> 00:24:02,160 decentralized poor network where we can use a network of miners 406 00:24:02,160 --> 00:24:04,800 or prover to generate proof of us to provide enough computation 407 00:24:04,800 --> 00:24:08,000 power for us to generate proof for large applications. 408 00:24:08,240 --> 00:24:11,680 So that's our initial idea firstly having this architecture 409 00:24:12,000 --> 00:24:14,840 and then the key VM, the key VM is just application. 410 00:24:15,120 --> 00:24:17,840 And then initially we are thinking about more even in our 411 00:24:17,840 --> 00:24:21,840 initial posting I I think in April 2021 is talking about like 412 00:24:21,840 --> 00:24:25,360 the key virtual machine how you the key CPU how that you know 413 00:24:25,360 --> 00:24:28,400 cycle works. But later I think I talked with 414 00:24:29,320 --> 00:24:33,240 I I met, I met Barry from ECM Foundation who is like who 415 00:24:33,440 --> 00:24:36,120 actually invented the key. Rob the the the, the concept of 416 00:24:36,160 --> 00:24:37,760 of the key. Robber Barry Whitehead, right, 417 00:24:37,920 --> 00:24:39,000 Yeah. Barry Whitehead, yeah. 418 00:24:39,280 --> 00:24:43,520 And he he, he has been thinking through like whether the QM is 419 00:24:43,520 --> 00:24:46,960 possible whether by code level like the QVM is possible or not. 420 00:24:47,240 --> 00:24:50,600 And then like we like he he, he read our documentation and then 421 00:24:50,600 --> 00:24:53,040 like say said OK so if you also want to build something like 422 00:24:53,040 --> 00:24:56,480 this we also had some idea why not like you know we collaborate 423 00:24:56,480 --> 00:24:59,240 on this and then like share if you think this is the right 424 00:24:59,240 --> 00:25:02,440 approach to to build why don't we just building the open 425 00:25:02,520 --> 00:25:05,480 everything is open source and then like just build that. 426 00:25:05,880 --> 00:25:10,040 I think at that time that's where I think all the the QVM 427 00:25:10,040 --> 00:25:13,480 even including like like Hermes, like there's the idea for from 428 00:25:13,480 --> 00:25:15,920 from Geordie like all start from there. 429 00:25:16,200 --> 00:25:18,840 So I think it's Barry's discussing some with someone 430 00:25:18,840 --> 00:25:22,280 else to about this idea for how you use lookup to solve the 431 00:25:22,280 --> 00:25:26,360 biggest problem for read and write to storage. 432 00:25:26,360 --> 00:25:31,320 Because because like is there not proof you to prove program 433 00:25:31,320 --> 00:25:34,160 for for some static program which for example you have some 434 00:25:34,160 --> 00:25:39,200 sorting, you prove this fixed program but VM needs to read and 435 00:25:39,200 --> 00:25:43,560 write the state, the storage and that mapping is very costly 436 00:25:43,560 --> 00:25:46,160 because you need to use a merkle tree every time you read from 437 00:25:46,160 --> 00:25:48,040 element you need to prove a merkle tree branch. 438 00:25:48,560 --> 00:25:52,320 And the idea is that we can use look up to solve this problem 439 00:25:52,480 --> 00:25:54,800 like totally like because you only need to build a look up 440 00:25:54,800 --> 00:25:58,480 table and then you can do efficient batch look up to into 441 00:25:58,480 --> 00:26:01,640 that that that table. And that's where all the ideas 442 00:26:01,640 --> 00:26:04,440 come from including like Jordis. I think Barry has been talking 443 00:26:04,440 --> 00:26:07,600 with a lot of teams like including like Jordy, us as well 444 00:26:07,640 --> 00:26:10,760 as some other layer two teams and that we are the first one to 445 00:26:10,760 --> 00:26:14,360 commit to world to building the open and we want to kind of 446 00:26:14,360 --> 00:26:20,800 build with with their team as as exploration to kind of because 447 00:26:20,840 --> 00:26:24,120 we believe that building the public is a one of the core 448 00:26:24,120 --> 00:26:26,720 spirit of of ECM community. Why it's so vibrant like because 449 00:26:26,720 --> 00:26:28,640 people can kind of contribute freely. 450 00:26:28,640 --> 00:26:32,560 People can kind of collaborate freely And then I think Jordy is 451 00:26:32,560 --> 00:26:36,160 more leaning towards like stark based approach but the 452 00:26:36,160 --> 00:26:37,520 architecture is very very similar. 453 00:26:37,520 --> 00:26:42,280 It's all originally from the the like the the very like old doc 454 00:26:42,280 --> 00:26:45,600 where you use look up to solve this problem and then like he's 455 00:26:45,840 --> 00:26:49,120 he like I think then like he's back out like he want to use 456 00:26:49,120 --> 00:26:52,240 Stark he takes some different design choices and then like he 457 00:26:52,240 --> 00:26:55,320 choose to like build build with this this Polygon and build 458 00:26:55,320 --> 00:26:59,640 their own like I think he he builds their own like 459 00:27:00,080 --> 00:27:03,280 instruction sets to and also interpreter to interpret by code 460 00:27:03,320 --> 00:27:06,840 into that instruction set. But we are more firmly believe 461 00:27:06,840 --> 00:27:10,600 that directly build circuit for each op code so that you can 462 00:27:10,600 --> 00:27:12,840 have per op code mapping. You don't need to build any 463 00:27:12,840 --> 00:27:16,560 interpreter you don't need to build a a sequencer from 464 00:27:16,560 --> 00:27:18,760 scratch. You can reuse everything that 465 00:27:18,800 --> 00:27:21,680 ECM has. So that's why I like us and the 466 00:27:21,680 --> 00:27:24,880 PSE team which is short for privacy and scaling exploration 467 00:27:24,880 --> 00:27:29,400 team led by Barry Redhead and work towards the direction where 468 00:27:29,400 --> 00:27:33,560 it's OP code level compatible maximum the usability of there 1 469 00:27:33,560 --> 00:27:38,720 sequencer and then Jordy leads towards a stock based back end 470 00:27:38,920 --> 00:27:41,800 and some kind of other instruction set with an 471 00:27:41,800 --> 00:27:45,320 interpreter but still like with this kind of by code level 472 00:27:45,320 --> 00:27:47,840 compatibility. So that's where different 473 00:27:47,840 --> 00:27:50,960 approaches differ but it's all actually arranged from the 474 00:27:51,360 --> 00:27:55,000 breakthrough or the idea for OK use look up to do this use 475 00:27:55,000 --> 00:27:58,160 custom gate to kind of express your your op codes. 476 00:27:58,440 --> 00:28:01,080 So it's very similar. It's I think it's exactly the 477 00:28:01,080 --> 00:28:04,960 same timeline actually like because I I remember like you 478 00:28:04,960 --> 00:28:07,400 know Barry's talking with with many teams about you know 479 00:28:07,400 --> 00:28:10,200 whether they are interested in kind of building this together. 480 00:28:10,760 --> 00:28:13,200 But there will be different considerations behind teams 481 00:28:13,200 --> 00:28:16,960 around like how this need to be built and like different design 482 00:28:16,960 --> 00:28:19,160 choices. And we are more aligned on the 483 00:28:19,160 --> 00:28:22,880 other side, like, you know, use this kind of KDG, use snark, use 484 00:28:24,080 --> 00:28:27,280 maximum usability and build something like in the public. 485 00:28:27,560 --> 00:28:30,920 So yeah, that's the difference. But everyone starts the same 486 00:28:30,920 --> 00:28:33,440 point with a similar architecture for how you handle 487 00:28:33,440 --> 00:28:37,400 this memory thing. Yeah, this is super interesting 488 00:28:37,400 --> 00:28:40,480 to kind of fear that it kind of all started with Barry 489 00:28:40,480 --> 00:28:42,800 Whitehead's idea of kind of how to do the look up. 490 00:28:44,000 --> 00:28:47,040 So you guys recently kicked off your mainnet. 491 00:28:47,160 --> 00:28:50,000 If I remember correctly, it was mid-october. 492 00:28:50,720 --> 00:28:54,960 So tell us about that. Yeah, it has been a Really. 493 00:28:54,960 --> 00:28:59,320 Really long journey from from search to to mid. 494 00:28:59,880 --> 00:29:03,400 I think we we we officially launched the the genesis happens 495 00:29:03,520 --> 00:29:09,360 in like October 10th and we we encode some like the future is 496 00:29:09,360 --> 00:29:14,600 open using like a word language to to to try that you know we 497 00:29:14,600 --> 00:29:18,440 think the word need to be connected it's global and the 498 00:29:18,440 --> 00:29:22,640 future is open and then like I think the official opening is 499 00:29:22,920 --> 00:29:28,960 mid-october I think before it's it's really exciting like the 500 00:29:28,960 --> 00:29:32,760 whole team is in Vietnam doing some offsites like we we 501 00:29:32,760 --> 00:29:35,440 actually they're like celebrating in person with each 502 00:29:35,440 --> 00:29:38,360 other like because it's a joint effort not only like research 503 00:29:38,360 --> 00:29:41,840 not only engineering but also like the whole team in like PD 504 00:29:41,840 --> 00:29:46,440 Devro like OPS all those different teams like joint for 505 00:29:46,440 --> 00:29:50,360 us to to build this together not only a product but also like not 506 00:29:50,360 --> 00:29:53,760 only a like just engineer effort but but it's a it's a community 507 00:29:53,920 --> 00:29:57,720 effort. I think we we have been like 508 00:29:57,800 --> 00:30:01,240 really doing a lot of lot of like preparation to make this 509 00:30:01,240 --> 00:30:04,040 happen like our test net have been running for over like 15 510 00:30:04,040 --> 00:30:07,320 months from pre alpha test net to alpha test net to beta test 511 00:30:07,320 --> 00:30:11,480 net lot of upgrades a lot of testing I think we like before 512 00:30:11,800 --> 00:30:14,960 before man launch we are very nervous about you know the 513 00:30:14,960 --> 00:30:17,920 security because you are really launching something play with 514 00:30:17,920 --> 00:30:21,240 like users assets like users can really deposit our money on your 515 00:30:21,240 --> 00:30:22,840 chain you really need to be responsible. 516 00:30:22,840 --> 00:30:28,680 So we we pay huge attention to like security monitoring system 517 00:30:28,960 --> 00:30:31,800 everything else to to kind of make sure it's a successful 518 00:30:31,800 --> 00:30:34,600 launch we have done like for example we we even like before 519 00:30:34,600 --> 00:30:37,720 we launched we fetched in our testing environment we fetched 520 00:30:37,840 --> 00:30:41,160 all the transactions from from Polygon from finance from from 521 00:30:41,160 --> 00:30:45,960 different chance that to kind of repay our in our testing 522 00:30:45,960 --> 00:30:48,480 environment to make sure that we can generate proof for those 523 00:30:48,480 --> 00:30:51,040 connections to kind of stress test. 524 00:30:51,680 --> 00:30:55,320 And also like we we we spend a quite amount of money on 525 00:30:55,320 --> 00:30:57,160 auditing like internal and external. 526 00:30:57,160 --> 00:31:00,520 Internally we have some kind of red team to looking keep looking 527 00:31:00,520 --> 00:31:05,760 for bugs attacking our our test net and the environment and how 528 00:31:05,760 --> 00:31:08,720 to make thick proof how to kind of attack the system. 529 00:31:08,920 --> 00:31:13,320 And externally there is like we contract the best the word class 530 00:31:13,320 --> 00:31:17,040 level of like auditor like including like open zabling 531 00:31:17,320 --> 00:31:20,800 Zadig for for for node and and the contract auditing a lot of 532 00:31:20,960 --> 00:31:23,840 complaints like that. We also like set up like $1 533 00:31:23,840 --> 00:31:27,320 million back bounty to to for for for the community to support 534 00:31:27,360 --> 00:31:30,560 any bugs in our in our system. So it's like a lot of 535 00:31:30,720 --> 00:31:35,760 preparation and also like we our team has been like you know 536 00:31:35,760 --> 00:31:38,640 talking is a lot of project to to deploy and test their 537 00:31:38,640 --> 00:31:41,840 application on Sympolia. Yeah I think I think it's all 538 00:31:41,880 --> 00:31:46,200 like it's very exciting moment that you turn you literally turn 539 00:31:46,200 --> 00:31:52,080 some open source project, open source demo to a product like 540 00:31:52,400 --> 00:31:56,960 like like 100% like ready products online. 541 00:31:57,240 --> 00:32:01,120 So I think that's a very very like incredible journey for me 542 00:32:01,120 --> 00:32:05,520 and also like you see like your your research results from from 543 00:32:05,520 --> 00:32:08,600 paper to a product and used by a lot of people. 544 00:32:08,880 --> 00:32:11,640 So it's it's it's quite exciting and we are really looking 545 00:32:11,640 --> 00:32:14,560 forward to kind of things after like you know how we build the 546 00:32:14,560 --> 00:32:16,480 community how we keep improving attack. 547 00:32:17,680 --> 00:32:20,960 Yeah, I think it's yeah it's it's really a moment that we we 548 00:32:20,960 --> 00:32:23,720 can't we we will remember like forever like you know it's 549 00:32:23,720 --> 00:32:27,480 launched and very exciting. Yeah, congratulations on that. 550 00:32:28,120 --> 00:32:30,800 Tell us about the Scroll community. 551 00:32:30,800 --> 00:32:34,760 So kind of who is your community and do you have people who are 552 00:32:34,760 --> 00:32:37,520 building on Scroller exclusively? 553 00:32:38,480 --> 00:32:42,240 Yeah, yeah, that makes sense. So first thing I think my 554 00:32:42,240 --> 00:32:45,360 definition for community is that there will be like engineer and 555 00:32:45,360 --> 00:32:48,200 researchers who are part of like they can build their own 556 00:32:48,200 --> 00:32:50,840 product, they can build their own project, they can do their 557 00:32:50,840 --> 00:32:53,280 own research. But we we communicate, we 558 00:32:53,280 --> 00:32:56,120 collaborate, we talk in the open, we call build some kind of 559 00:32:56,160 --> 00:32:57,720 talk about some some benchmark stuff. 560 00:32:58,040 --> 00:33:02,040 So those are I think are part of our community And by for example 561 00:33:02,040 --> 00:33:06,480 we are hosting ZK symposium which is like like due to the 562 00:33:06,480 --> 00:33:09,160 money that is currently recently reducing frequency a little bit 563 00:33:09,160 --> 00:33:11,800 but we will catch up very soon is that previously it's like 564 00:33:12,600 --> 00:33:17,520 biweekly cadence where we invite the ZK researchers and 565 00:33:17,520 --> 00:33:21,600 application developer to talk about their idea for building ZK 566 00:33:22,440 --> 00:33:24,280 like what application they are building using is AK. 567 00:33:24,280 --> 00:33:27,440 Because I think a lot of if if you look at Twitter like all the 568 00:33:27,440 --> 00:33:31,760 kind of places where people only like talk about like some 569 00:33:31,760 --> 00:33:34,160 marketing material. We'll talk about how great they 570 00:33:34,160 --> 00:33:37,920 are like how how fast they are like why they can achieve like 571 00:33:37,920 --> 00:33:40,280 you know 100% privacy all those stuff. 572 00:33:40,440 --> 00:33:43,440 But no one really dig into details about the architecture 573 00:33:43,720 --> 00:33:48,160 and want to build a like a place where people can like dive 574 00:33:48,320 --> 00:33:51,760 really, really deep into like what they are building and the 575 00:33:51,760 --> 00:33:54,240 architecture stuff. So I think it more happens in 576 00:33:54,240 --> 00:33:58,760 like Zika study club where people dive into like like very 577 00:33:58,760 --> 00:34:01,080 academic papers, some mathematical constructions. 578 00:34:01,200 --> 00:34:04,480 But very few people have some place to share very deeply about 579 00:34:04,480 --> 00:34:07,080 Zika application ideas like how they build this app, Zika 580 00:34:07,080 --> 00:34:09,639 application, what technology they are using, what the 581 00:34:09,639 --> 00:34:11,920 architecture look like. And then that's where Zika 582 00:34:11,920 --> 00:34:16,800 symbolism sit where people have one hour time to present their 583 00:34:16,840 --> 00:34:20,080 idea, be in depth about what applications they are building 584 00:34:20,080 --> 00:34:23,360 the technology and then people can like ask questions. 585 00:34:23,679 --> 00:34:26,040 So I think that those are like in that way we are building our 586 00:34:26,040 --> 00:34:29,679 like research community and also because everything is building 587 00:34:29,679 --> 00:34:33,159 the open like as I mentioned like our our the key VM reuse 588 00:34:33,159 --> 00:34:35,600 all the toolings from for example like the pre library 589 00:34:35,600 --> 00:34:38,400 from from the cash team. And there are a lot of other 590 00:34:38,400 --> 00:34:42,679 teams like also building on top of this this this crypto library 591 00:34:42,679 --> 00:34:46,320 that those are also I as part of our community like the key 592 00:34:46,320 --> 00:34:50,400 community to to build that's what one part and second part is 593 00:34:50,400 --> 00:34:52,960 more like developer community where people deploy the 594 00:34:53,159 --> 00:34:56,760 applications and we introduce after after Minet we introduced 595 00:34:56,840 --> 00:34:59,040 something called scroll or Ranger NFT. 596 00:34:59,600 --> 00:35:03,200 So you really I'm not a big fan of of of this kind of NFT thing 597 00:35:03,360 --> 00:35:06,880 but it's actually a SOAP bound where if you deploy applications 598 00:35:07,360 --> 00:35:11,080 you you you are early in our ecosystem you you can claim 599 00:35:11,080 --> 00:35:14,320 you're like you know one really cool scroll or Ranger NFT from 600 00:35:14,560 --> 00:35:19,120 from us I think that's design. I don't know if you have noticed 601 00:35:19,120 --> 00:35:22,640 that it's totally different from like just images like you really 602 00:35:23,040 --> 00:35:26,240 have to just images stored somewhere else and then you put 603 00:35:26,240 --> 00:35:30,360 some kind of link on on chain which is a a fake FT What we do 604 00:35:30,360 --> 00:35:33,480 is that we do something similar to like Unicorn with three which 605 00:35:33,480 --> 00:35:37,480 is a generative curve. It's a polynomial where if you 606 00:35:37,480 --> 00:35:41,040 deploy a contract you pick because the time you deployed 607 00:35:41,040 --> 00:35:44,360 because of the kind of the sequence like how many contract 608 00:35:44,360 --> 00:35:48,360 you are deployed. Then like you can get a 609 00:35:48,360 --> 00:35:51,720 different polynomial and then like we can draw this polynomial 610 00:35:52,440 --> 00:35:56,920 like on chain like you know you get a different like I think if 611 00:35:56,920 --> 00:36:00,560 you deploy earlier that the the degree will be four or five I I 612 00:36:00,560 --> 00:36:03,200 can't remember. So it's like you get this kind 613 00:36:03,200 --> 00:36:06,560 of curve with with different like so many kind of turning 614 00:36:06,560 --> 00:36:11,320 points and to show that you are early in our ecosystem and we 615 00:36:11,320 --> 00:36:14,440 really appreciate our effort because you know being in a 616 00:36:14,480 --> 00:36:18,160 ecosystem early it definitely introduced some risk for OK, so 617 00:36:18,160 --> 00:36:20,920 whether it's better tested or not so we will come the early 618 00:36:20,920 --> 00:36:24,000 builders to to join our ecosystem that's only for 619 00:36:24,000 --> 00:36:27,320 welcome and like to prove that you are you are there and you 620 00:36:27,320 --> 00:36:30,200 are there earlier. And so like everything we do is 621 00:36:30,200 --> 00:36:33,840 like I think it's very cool and very it's based on like 622 00:36:33,840 --> 00:36:36,840 fundamentals that make as possible which is 0 know proof 623 00:36:36,840 --> 00:36:40,240 and polynomials and then like you know application you deploy 624 00:36:40,280 --> 00:36:43,800 you got a polynomial you got a NFT like you know which draw 625 00:36:43,800 --> 00:36:45,640 this kind of really cool polynomial. 626 00:36:46,080 --> 00:36:48,120 I think that that that will definitely introduce some kind 627 00:36:48,120 --> 00:36:51,640 of a lot of people are just trying to trying to kind of 628 00:36:51,640 --> 00:36:55,640 deploy maybe a bunch of like ERC 20 contracts but that's within 629 00:36:55,640 --> 00:37:00,240 our prediction our our kind of motivation objective is that 630 00:37:00,440 --> 00:37:04,200 encourage more people to try deploy your first first contract 631 00:37:04,680 --> 00:37:07,880 and in a neutral way it's not like we select or we give you 632 00:37:07,880 --> 00:37:11,160 this we give you that in a unfair way but we we are more we 633 00:37:11,160 --> 00:37:13,840 are we we want to hold this kind of like a blockchain's principle 634 00:37:13,840 --> 00:37:16,480 of being credibly neutral and then people every people you 635 00:37:16,480 --> 00:37:20,920 deploy you you can get something as and also experience like how 636 00:37:20,920 --> 00:37:23,760 seamless that is. So I think that's that's 637 00:37:23,760 --> 00:37:27,120 something like which there are some community from there and 638 00:37:27,120 --> 00:37:30,360 also like there are some more native applications which we 639 00:37:30,360 --> 00:37:32,960 share the same value. And then they believe that you 640 00:37:32,960 --> 00:37:36,240 know we believe our value of being open, being commit driven, 641 00:37:36,240 --> 00:37:38,600 being credibly neutral. And then they come to us because 642 00:37:38,640 --> 00:37:41,960 we we we don't provide brand to applications specifically 643 00:37:41,960 --> 00:37:45,520 because we think that will make your ecosystem become a 0 sum 644 00:37:45,520 --> 00:37:48,560 game because basically like there will always be new chance 645 00:37:48,560 --> 00:37:50,240 like new layer one, new layer two. 646 00:37:50,240 --> 00:37:53,400 They always launch their, their token with they they raise a 647 00:37:53,400 --> 00:37:56,400 lot, much of money. They can give a lot to, to a new 648 00:37:56,400 --> 00:37:58,120 developer. Like I give you this money you 649 00:37:58,120 --> 00:38:01,160 come to our chain to deploy but then like it will become a 650 00:38:01,160 --> 00:38:05,280 competing game which I give you this amount the other chain give 651 00:38:05,280 --> 00:38:07,560 you that amount and then you go to the other chain. 652 00:38:07,880 --> 00:38:11,720 And it seems like it's not like we are we are making the whole 653 00:38:11,720 --> 00:38:14,760 ecosystem grow. We are like making this kind of 654 00:38:14,760 --> 00:38:19,160 small pie for crypto larger but it's more like I want to compete 655 00:38:19,160 --> 00:38:21,720 with with with each other. So what we want to do is that we 656 00:38:21,720 --> 00:38:23,760 want to attract. We want to focus on like 657 00:38:23,800 --> 00:38:27,560 building stuff and attract more organic community where you know 658 00:38:27,560 --> 00:38:30,320 they are very organic, attract the best because they see like 659 00:38:30,320 --> 00:38:35,240 there is a huge opportunity I'll scroll by being early and and 660 00:38:35,240 --> 00:38:38,120 also like it's one of the most legit and general purpose like 661 00:38:38,120 --> 00:38:41,000 top top layer two and there will be a huge opportunity there. 662 00:38:41,320 --> 00:38:44,160 So what we are focusing on is that we introduce we we make we 663 00:38:44,160 --> 00:38:46,840 get everything ready for developers for example like we 664 00:38:46,840 --> 00:38:50,600 have ether scan support so users and develop can use your most 665 00:38:50,600 --> 00:38:53,400 familiar infrastructure. We get chat link as Oracle 666 00:38:53,400 --> 00:38:57,320 support which means like you know all the kind of defy can 667 00:38:57,320 --> 00:38:59,480 use your your most familiar Oracle. 668 00:38:59,480 --> 00:39:03,040 We get we get all the kind of indexing RPC provider ready. 669 00:39:03,200 --> 00:39:06,320 So that's something we will focus on like build a 670 00:39:06,320 --> 00:39:09,840 environment that developer are easy to build but we don't like 671 00:39:09,840 --> 00:39:12,240 enforce you to to build like what we want. 672 00:39:12,240 --> 00:39:14,200 We we just create this kind of foundation. 673 00:39:14,200 --> 00:39:17,520 We keep improve our documentation tutorial workshops 674 00:39:18,280 --> 00:39:20,400 and then like a lot of people are actually attracted by this 675 00:39:20,400 --> 00:39:23,760 it's like for some more like commercialized application for 676 00:39:23,760 --> 00:39:25,920 them. One is there is versa like you 677 00:39:25,920 --> 00:39:29,560 know there are some other one is which are native to school and 678 00:39:29,560 --> 00:39:32,840 there are some like new D file like coke finance and some other 679 00:39:33,280 --> 00:39:34,560 financial applications on scroll. 680 00:39:34,560 --> 00:39:38,320 And amazingly it's like a lot of like small AI games like 681 00:39:38,320 --> 00:39:41,040 happened to our scroll like I think days ago like I don't know 682 00:39:41,040 --> 00:39:44,040 who build this but it's like there's a game called chat MPC 683 00:39:44,040 --> 00:39:49,240 where you you chat with with the AI image and then like you could 684 00:39:49,240 --> 00:39:52,200 try to negotiate through this conversation about how much you 685 00:39:52,200 --> 00:39:56,160 can get the and then like you end up if you end up being being 686 00:39:56,160 --> 00:39:59,040 high you can get some score. It's it's very simple game but 687 00:39:59,040 --> 00:40:01,920 it's very fun. So I think being fun is really 688 00:40:01,920 --> 00:40:05,840 one important. Way to attract attract builders 689 00:40:06,160 --> 00:40:09,120 I I think a lot of I know like a lot of like applications are 690 00:40:09,120 --> 00:40:12,880 still building the because the more serious and more like 691 00:40:14,120 --> 00:40:17,360 commercialize more ready the project is it will take a longer 692 00:40:17,360 --> 00:40:18,920 time to to build and being ready. 693 00:40:19,160 --> 00:40:22,160 But for for every laugh with listening like pay attention to 694 00:40:22,160 --> 00:40:25,720 our ecosystem account our main account you will notice a lot of 695 00:40:25,720 --> 00:40:28,320 good opportunities good projects at building on top of us which 696 00:40:28,320 --> 00:40:31,960 will come out in the next few months and in terms of like also 697 00:40:31,960 --> 00:40:33,640 support a lot of grassroots hexons. 698 00:40:33,920 --> 00:40:37,920 We have been I I think we we probably participant over like 699 00:40:37,920 --> 00:40:42,200 20 hexons maybe like around the world like everywhere almost and 700 00:40:42,200 --> 00:40:46,560 then like we a very recent one is at East Global in New York. 701 00:40:47,000 --> 00:40:50,640 We it's the first time that we surpassed all the all other 702 00:40:50,640 --> 00:40:56,040 chance in term of deployment and with small price. 703 00:40:56,240 --> 00:40:58,280 So you already like you know it's it's not just you know 704 00:40:58,800 --> 00:41:00,760 people offer high price or people deploy on you. 705 00:41:01,000 --> 00:41:05,600 But if I realize EVM and then like you can provide something 706 00:41:05,600 --> 00:41:10,760 kind of subtle changes or like something cool like you got the 707 00:41:10,760 --> 00:41:14,120 most deployment And then like it's very welcome in this kind 708 00:41:14,120 --> 00:41:18,120 of hacker group And it's very exciting because it's even like 709 00:41:18,120 --> 00:41:21,840 before our minute launch where a lot of other chance have all 710 00:41:21,840 --> 00:41:24,760 those kind of even better infrastructure support but we 711 00:41:24,760 --> 00:41:28,200 can still we get so many hackers interesting kind of building on 712 00:41:28,200 --> 00:41:29,760 top of us. I think it's definitely 713 00:41:29,760 --> 00:41:34,080 something like which we are kind of celebrating like because this 714 00:41:34,080 --> 00:41:36,960 is very organic grassroots hexon project. 715 00:41:37,320 --> 00:41:40,680 I do believe that in this way like you can you can grow your 716 00:41:40,680 --> 00:41:42,960 grassroots community to a large extent. 717 00:41:43,320 --> 00:41:46,040 So those are like the all developer communities we have, 718 00:41:46,280 --> 00:41:49,080 you know including like as I mentioned like some wallets, 719 00:41:49,080 --> 00:41:53,600 some some games, some some kind of like DF applications and 720 00:41:53,640 --> 00:41:57,760 there's a lot of Hexon groups. And the other other step, I 721 00:41:57,760 --> 00:42:01,520 think looking forward we will expect more regional community. 722 00:42:01,680 --> 00:42:05,360 So we will start seriously building community in several 723 00:42:05,360 --> 00:42:13,000 places including like Turkey, Nigeria, Malaysia and Argentina. 724 00:42:13,000 --> 00:42:15,120 That's our our, our plan for like we are. 725 00:42:15,120 --> 00:42:18,000 We seriously want to put some efforty to grow the community 726 00:42:18,000 --> 00:42:23,200 there because our our our mission is really to scale to 727 00:42:23,200 --> 00:42:26,720 scale Etherium to sometimes like you know if you have like we 728 00:42:26,720 --> 00:42:31,440 want to codify trust and like empower this ownership for for 729 00:42:31,440 --> 00:42:33,360 individual and chief financial inclusion. 730 00:42:33,560 --> 00:42:36,720 So and you see that all the blockchains are not scalable 731 00:42:36,720 --> 00:42:41,080 enough for BF users for even like they they don't even know 732 00:42:41,080 --> 00:42:43,320 like where all those run BI users come from. 733 00:42:44,000 --> 00:42:48,920 So we have very clear goal for how to guide those users because 734 00:42:49,640 --> 00:42:54,440 the the the users like we need crypto app where the 1 billion 735 00:42:54,440 --> 00:42:58,040 people really come from. And you do see like you know in 736 00:42:58,040 --> 00:43:01,640 US, in even in China, in some kind of other more like in 737 00:43:01,640 --> 00:43:04,880 Europe a lot of places people don't really need crypto and 738 00:43:05,400 --> 00:43:08,160 because they have really established like financial 739 00:43:08,600 --> 00:43:11,600 system they only do that for cross-border like you know 740 00:43:11,600 --> 00:43:13,360 payment and maybe people use that. 741 00:43:13,680 --> 00:43:16,680 But in a lot of places like like Africa like you know I have been 742 00:43:16,680 --> 00:43:20,240 there for for for 10 days for this year to trip and to really 743 00:43:20,240 --> 00:43:22,400 understand like how people are using crypto in practice. 744 00:43:22,800 --> 00:43:24,800 You figure out that a lot of people are really suffer from 745 00:43:24,800 --> 00:43:29,080 this kind of currency inflation and they have a lot of problems 746 00:43:29,080 --> 00:43:30,880 that can be tackled by blockchain. 747 00:43:31,280 --> 00:43:35,560 We really want to kind of help those users and onboard user use 748 00:43:35,560 --> 00:43:39,360 those users to ECM ecosystem. So that's why like we we 749 00:43:39,360 --> 00:43:43,080 selectively choose those places which they suffer from the the 750 00:43:43,080 --> 00:43:46,920 problems of and they have really need for crypto And then like we 751 00:43:46,920 --> 00:43:49,360 want to double down effort in building regional community. 752 00:43:49,840 --> 00:43:53,760 And eventually I think I I'm imagine like in a few few months 753 00:43:53,760 --> 00:43:57,320 or like a few years school will be think about as a the layer 754 00:43:57,320 --> 00:44:00,600 tool that has the most real users coming from developing 755 00:44:00,600 --> 00:44:03,760 countries from from from from someplace in Asia some. 756 00:44:04,120 --> 00:44:08,560 So all those places where if application come to deploy they 757 00:44:08,560 --> 00:44:10,320 know that there is real user there. 758 00:44:10,320 --> 00:44:13,560 It's not just the same group of people like in migrating from 759 00:44:13,560 --> 00:44:15,320 this chain to that chain because it's a new chain. 760 00:44:15,720 --> 00:44:20,080 So that's our like mission and I I consider that to be part of 761 00:44:20,080 --> 00:44:22,920 our community but we are still building that we we just start 762 00:44:22,920 --> 00:44:26,680 on Turkey but we was like putting somewhere like. 763 00:44:27,560 --> 00:44:30,360 So if you look at like you know this year a lot of like project 764 00:44:30,360 --> 00:44:34,320 was out to from from from DEM connect but we are still like 765 00:44:34,320 --> 00:44:37,480 building a very very large probably largest ever event 766 00:44:37,600 --> 00:44:41,560 called layer today with layer to beat we are Co hosting that it's 767 00:44:41,720 --> 00:44:45,160 like over 2000 people capacity to to kind of talk about layer 768 00:44:45,160 --> 00:44:48,280 tooth and we are seriously building a commuting starting 769 00:44:48,280 --> 00:44:51,800 from Turkey but other places we will if you are like you know in 770 00:44:51,800 --> 00:44:54,920 the regions you really want to change change people's life 771 00:44:54,920 --> 00:44:57,720 there then like you know you can you can talk with us and we we 772 00:44:57,720 --> 00:45:00,400 we want to kind of build the committee there seriously we are 773 00:45:00,880 --> 00:45:04,000 I think I see like where our community can come from. 774 00:45:04,880 --> 00:45:06,480 OK. So basically what what you're 775 00:45:06,480 --> 00:45:09,800 saying is you're trying to grow the community organically 776 00:45:10,480 --> 00:45:13,920 through culture rather than through kind of like monetary 777 00:45:13,920 --> 00:45:17,080 incentives, which is kind of often the way that it goes in 778 00:45:17,080 --> 00:45:20,520 crypto. Speaking of incentives, what's 779 00:45:20,520 --> 00:45:25,880 this, the token situation First role, Yeah, it's for our legal 780 00:45:25,880 --> 00:45:29,080 reason like you know, we we we haven't like you know, really 781 00:45:29,080 --> 00:45:32,560 publicly. You know, I'm more thinking like 782 00:45:32,560 --> 00:45:36,160 you know, it's more like a mechanism We are thinking about 783 00:45:36,160 --> 00:45:38,000 like why you, why you even need this. 784 00:45:38,000 --> 00:45:40,520 Like you know, maybe there are some reasons for decided 785 00:45:40,520 --> 00:45:44,000 sequencer disapprover, but there might be some other other other 786 00:45:44,000 --> 00:45:48,360 models for that. So I think we asked you thinking 787 00:45:48,360 --> 00:45:53,920 through what's the best mechanism and and like incentive 788 00:45:53,920 --> 00:45:58,280 a model for for our system to work the best instead of OK so 789 00:45:58,360 --> 00:46:02,280 just this this this token thing. But yeah we we we have been 790 00:46:02,280 --> 00:46:06,920 working on some mechanism design which aligned with our system to 791 00:46:06,920 --> 00:46:09,160 make our system become more stable and more useable. 792 00:46:09,680 --> 00:46:12,880 But you know for some legal or some some other concerns like 793 00:46:12,880 --> 00:46:16,000 you know we are not yet. But but currently. 794 00:46:16,520 --> 00:46:19,600 Yes on scroll is paid in East, yes yes. 795 00:46:21,040 --> 00:46:24,120 You already alluded to it in your last answer. 796 00:46:24,760 --> 00:46:27,240 You currently still have centralized sequences. 797 00:46:27,680 --> 00:46:30,600 It's kind of it's been normalized over the last like 798 00:46:30,600 --> 00:46:33,640 year or two. But in principle, what it means 799 00:46:33,640 --> 00:46:38,080 is that you kind of you have one or several permissioned 800 00:46:38,800 --> 00:46:42,800 sequences, people who can, or entities who can actually build 801 00:46:42,800 --> 00:46:47,960 blocks, which is very much not. What block chains should strive 802 00:46:47,960 --> 00:46:50,840 to because kind of in the in terms of decentralized 803 00:46:51,120 --> 00:46:54,920 decentralization, kind of your least decentralized component 804 00:46:54,920 --> 00:46:57,840 kind of determines how decentralized like the entire 805 00:46:57,840 --> 00:47:00,160 system, right. And so if you have one 806 00:47:00,400 --> 00:47:05,280 centralized sequencer, you have essentially the entire chain is 807 00:47:05,280 --> 00:47:09,040 built by a single entity. What, what are your plans 808 00:47:09,040 --> 00:47:12,960 towards decentralization? Yeah, that's a great question. 809 00:47:13,320 --> 00:47:17,400 So I definitely agree that you know that's in some sense like 810 00:47:17,400 --> 00:47:20,440 it's different from like whole. Normally like a layer, layer one 811 00:47:20,440 --> 00:47:24,080 blockchain works and like why only a few parties can can 812 00:47:24,080 --> 00:47:27,400 produce block. But I think when I think about 813 00:47:27,400 --> 00:47:29,760 this desensitization problem, I I think about like you know 814 00:47:30,360 --> 00:47:32,200 again like from first things for like why you need 815 00:47:32,200 --> 00:47:34,520 desensitization. And there are several reasons 816 00:47:34,520 --> 00:47:37,680 like in layer 2 contexts specifically they are like maybe 817 00:47:37,800 --> 00:47:41,160 maybe 3 aspects. I I think one is sequencer who 818 00:47:41,160 --> 00:47:44,080 generated block for you. So who can you know produce a 819 00:47:44,080 --> 00:47:46,640 block and like then pass that prover. 820 00:47:46,960 --> 00:47:49,960 Prover is basically generated. They keep proof for each block 821 00:47:50,240 --> 00:47:53,200 and to prove that OK so all the transactions at this block is 822 00:47:53,200 --> 00:47:56,840 valid and then like finally some in this, this, this block and 823 00:47:56,840 --> 00:47:58,640 the proof on chain and then on chain. 824 00:47:58,640 --> 00:48:01,200 Verifier will verify this. So there are two parties, one is 825 00:48:01,200 --> 00:48:04,520 sequencer generate block, one is prover like generate proofs. 826 00:48:04,760 --> 00:48:08,840 And two parts definitely need to be decentralized but for 827 00:48:08,840 --> 00:48:11,640 different reasons. So for for a sequencer the 828 00:48:11,640 --> 00:48:15,120 reason is that so actually like one thing I can just to to add 829 00:48:15,120 --> 00:48:18,240 to this is that even if we are using a centralized sequencer 830 00:48:18,760 --> 00:48:22,960 like what every layer twist is doing is that it doesn't 831 00:48:22,960 --> 00:48:25,400 influence your users fund security. 832 00:48:25,720 --> 00:48:29,600 Because if you think about like So what what the bad things can 833 00:48:29,600 --> 00:48:31,240 do like from a single perspective. 834 00:48:31,560 --> 00:48:34,240 So you can set a transaction sequencer includes that 835 00:48:34,240 --> 00:48:37,000 transaction block, prove our genet block, genet proof for 836 00:48:37,000 --> 00:48:39,480 this block. So if if we are operating a 837 00:48:39,480 --> 00:48:41,640 sequencer we want to do something bad right. 838 00:48:41,760 --> 00:48:45,240 If we insert a bad transaction then we can't generate proof for 839 00:48:45,240 --> 00:48:47,800 this bad transaction. Because you know the case the 840 00:48:47,800 --> 00:48:52,440 key proof, well kind of only attached to like the right 841 00:48:52,960 --> 00:48:54,280 transaction, the valid transaction. 842 00:48:54,280 --> 00:48:56,600 So we can't insert any bad bad transaction. 843 00:48:56,600 --> 00:48:59,480 So that's number one because of the key proof you can't do 844 00:48:59,480 --> 00:49:01,840 something bad. Second thing is that what you 845 00:49:01,840 --> 00:49:03,680 can do is that you can censor transaction. 846 00:49:03,960 --> 00:49:07,000 If users send their transaction to a centralized sequencer, send 847 00:49:07,000 --> 00:49:10,760 your sequencer say no I want to include you and but that can be 848 00:49:11,160 --> 00:49:15,760 like you know solved by OK so if users find that one one layer to 849 00:49:15,760 --> 00:49:19,200 accept their transaction, it can send this transaction on chain 850 00:49:19,480 --> 00:49:23,200 and then like layer one like verifier where you force that if 851 00:49:23,200 --> 00:49:25,600 you don't include this transaction in maybe 20-4 days 852 00:49:25,720 --> 00:49:27,240 that anyone can send me the block. 853 00:49:27,840 --> 00:49:30,600 So in that way like you know you can't avoid this kind of 854 00:49:30,600 --> 00:49:35,480 censorship resistance problem like and and users will never 855 00:49:35,560 --> 00:49:38,080 suffer from the problem of you can't withdraw fund because you 856 00:49:38,080 --> 00:49:41,560 can always do that we can't you know even if you reject like you 857 00:49:41,600 --> 00:49:45,440 you can you can do that and then so let's go back to so 858 00:49:45,440 --> 00:49:47,920 sensorized sequence of that influence the the system's kind 859 00:49:47,920 --> 00:49:52,440 of security but what it really influence is that there are some 860 00:49:52,440 --> 00:49:54,960 censorship like real time censorship resistant problem. 861 00:49:55,200 --> 00:49:58,160 So imagine like you know you will be liquidated in your next 862 00:49:58,160 --> 00:50:01,920 one minute you send a transaction to to deposit some 863 00:50:01,920 --> 00:50:04,560 money, but that we reject and then you are liquidated. 864 00:50:04,800 --> 00:50:08,240 And then even if like your transaction is like guaranteed 865 00:50:08,240 --> 00:50:12,120 to be included in in next a few blocks, it doesn't really matter 866 00:50:12,120 --> 00:50:13,560 because you're already being liquidated. 867 00:50:13,960 --> 00:50:17,040 So that's a problem and that's where you can you decide 868 00:50:17,040 --> 00:50:19,840 sequencer to to to help. But again like there are still 869 00:50:19,840 --> 00:50:21,480 maybe some other way to solve this problem. 870 00:50:21,480 --> 00:50:24,200 Maybe you can encrypt your use some equipment pool where you 871 00:50:24,200 --> 00:50:26,760 increase the transaction where sequencer can distinguish which 872 00:50:26,760 --> 00:50:29,480 transaction which and that it had to include and then like 873 00:50:29,480 --> 00:50:32,800 after decryption, you know like what's what's he included that 874 00:50:32,800 --> 00:50:34,000 can also solve this problem right. 875 00:50:34,000 --> 00:50:39,080 So like it's decimation. Sequencer is a tool not like we 876 00:50:39,080 --> 00:50:43,280 have to kind of oh we we have to for decimation we decentralize 877 00:50:43,440 --> 00:50:46,480 but it's for solving some problems and we we do think 878 00:50:46,680 --> 00:50:49,880 that's a promising way to solve this real time censorship recent 879 00:50:49,880 --> 00:50:51,520 problem. And also there might be some 880 00:50:51,520 --> 00:50:54,880 legal legal issues where there are multiple like you know 881 00:50:54,880 --> 00:50:57,440 entities running sequencer in different regions than like 882 00:50:58,200 --> 00:51:01,600 there's a risk for like being more censorship reasons can be 883 00:51:01,600 --> 00:51:03,960 reduced. So that's for sequencer and for 884 00:51:03,960 --> 00:51:07,360 the prover there are some kind of problems The thing that you 885 00:51:07,360 --> 00:51:12,120 can solve is that it can it can be very scalable where imagine 886 00:51:12,120 --> 00:51:15,720 that there is a network of of of minor or or prover running 887 00:51:15,920 --> 00:51:18,040 running or kind of proving proving algorithm. 888 00:51:18,800 --> 00:51:21,600 So it's very different from proof of work, even if you also 889 00:51:21,600 --> 00:51:24,200 have a high requirement for hardware because they give proof 890 00:51:24,200 --> 00:51:27,600 is that like proof of work is basically like 10,000 people 891 00:51:27,600 --> 00:51:30,640 computing for the same randomness and then who get the 892 00:51:30,640 --> 00:51:33,960 answer who gets some meat. But the key proof is that you 893 00:51:33,960 --> 00:51:37,960 have a fixed deterministic algorithm and anyone can execute 894 00:51:38,320 --> 00:51:42,320 and get it and that's it. So it's it's it's quite 895 00:51:42,320 --> 00:51:44,360 different. And so in terms of it's more 896 00:51:44,360 --> 00:51:47,160 like outsource computation, outsource this useful 897 00:51:47,160 --> 00:51:50,640 computation to the provers. And then like having this 898 00:51:50,640 --> 00:51:54,400 essential prover network can help you to kind of having some 899 00:51:54,400 --> 00:51:59,080 backup if you know one prover party just like goes down and 900 00:51:59,080 --> 00:52:01,760 then like other other prover can still generate proof for you. 901 00:52:01,760 --> 00:52:04,560 So you have some backup, you have some more stronger liveness 902 00:52:04,560 --> 00:52:05,920 guarantee. And also you don't need to buy 903 00:52:05,920 --> 00:52:09,160 machines yourself, it's more scalable, like people can use 904 00:52:09,160 --> 00:52:10,400 their own machine to generate proof. 905 00:52:11,560 --> 00:52:17,000 Yeah, it's more resilient. So kind of to to actually gain 906 00:52:17,200 --> 00:52:20,080 the forced inclusion that you talked about earlier if you're 907 00:52:20,080 --> 00:52:25,000 being censored on on the A2? You need to make all transaction 908 00:52:25,000 --> 00:52:29,800 data available on the L1. So basically this is kind of why 909 00:52:29,800 --> 00:52:33,640 we're now getting Deng sharding with blobs, where kind of data 910 00:52:34,440 --> 00:52:38,080 is going to get where temporary data is going to get much 911 00:52:38,080 --> 00:52:42,720 cheaper and so on. But it's still the main cost 912 00:52:42,720 --> 00:52:47,720 driver for L twos. So it's on the order of 95% or 913 00:52:47,720 --> 00:52:50,200 upwards of that as compared to kind of just? 914 00:52:51,280 --> 00:52:53,480 The check insurance that you kind of need to do to kind of 915 00:52:53,480 --> 00:52:57,120 prove the state right. So there's a couple of L twos 916 00:52:57,640 --> 00:53:01,520 that kind of have gone the villidium route recently where 917 00:53:01,520 --> 00:53:04,520 basically they decided to not actually post all transaction 918 00:53:04,520 --> 00:53:10,160 data to L1 just because it kind of drives up price on the L2. 919 00:53:11,040 --> 00:53:16,040 Currently prices aren't crazy, but I mean they have been crazy 920 00:53:16,040 --> 00:53:17,840 in the past. They have probably become crazy 921 00:53:17,840 --> 00:53:19,960 again at some point, which were kind of mean. 922 00:53:20,400 --> 00:53:24,000 That lots of applications that are currently viable, kind of 923 00:53:24,000 --> 00:53:28,480 like the AI game where you kind of you negotiate with an NPC, 924 00:53:29,000 --> 00:53:31,920 they will no longer be viable on L2, right? 925 00:53:32,160 --> 00:53:35,920 So what's what's your strategy there? 926 00:53:36,440 --> 00:53:40,240 So I mean basically even when den clotting comes, call data 927 00:53:40,240 --> 00:53:43,480 will become maybe 10 times cheaper or so, but this is still 928 00:53:43,480 --> 00:53:45,680 pretty expensive for lots of applications. 929 00:53:45,680 --> 00:53:48,480 So what are your thoughts on kind of going the Validium 930 00:53:48,480 --> 00:53:51,000 route. Yeah that's a that's a great 931 00:53:51,000 --> 00:53:53,840 question. So I I think for this with Idi 932 00:53:53,840 --> 00:53:57,120 think so I I think with DDM definitely trade off some 933 00:53:57,120 --> 00:54:01,360 security for being cheaper. But for for us I think currently 934 00:54:01,360 --> 00:54:04,520 we are sticking to I I think schools mentioned we'll stick to 935 00:54:04,520 --> 00:54:08,760 posting data on ECM and strictly you hide the same security from 936 00:54:08,760 --> 00:54:11,880 ECM. So because because our I I think 937 00:54:11,880 --> 00:54:14,520 for most of the row up, like different row up can have 938 00:54:14,520 --> 00:54:17,680 different value proposition. I think for us it's always like 939 00:54:17,680 --> 00:54:21,200 security first is always like drive most of our design 940 00:54:21,200 --> 00:54:24,000 decisions including like that's why we we do this auditing this 941 00:54:24,000 --> 00:54:25,640 way. We open sourcing this way. 942 00:54:25,640 --> 00:54:28,800 We have backbounding this way. So it drives almost all the 943 00:54:28,800 --> 00:54:33,360 decisions because we do feel like there will be a a crucial 944 00:54:33,360 --> 00:54:37,040 amount of applications that need this level of guarantee and then 945 00:54:37,040 --> 00:54:39,280 like we we we make this become our first priority. 946 00:54:39,280 --> 00:54:41,760 So for a lot for for quite a long time we will stick to this 947 00:54:42,320 --> 00:54:47,480 like approach and get the what we what we think is that I think 948 00:54:47,480 --> 00:54:49,840 we're telling you recently just post A blog post about like 949 00:54:49,840 --> 00:54:53,760 different layer tooth and there is actually a spectrum for what 950 00:54:53,760 --> 00:54:59,160 applications need like on the on the kind of left hand side 951 00:54:59,160 --> 00:55:03,120 there's a key store you know which is the the fundamentals of 952 00:55:03,160 --> 00:55:05,760 of all the smart contract wallets storing their keys key 953 00:55:05,760 --> 00:55:08,920 mappings, key value mappings which need extremely high 954 00:55:08,920 --> 00:55:13,800 security and then there's ENS which store your identity 955 00:55:13,800 --> 00:55:16,280 information. And then the more on the left 956 00:55:16,280 --> 00:55:20,040 side the the higher security you want like maybe C5 institutional 957 00:55:20,040 --> 00:55:23,040 money and some government we they want to kind of issue some 958 00:55:23,080 --> 00:55:26,320 something important they need security on the right hand side 959 00:55:26,320 --> 00:55:30,400 it might be gaming like some NFT or some kind of those things. 960 00:55:30,600 --> 00:55:33,960 I think our value polarization is that score man chain will lie 961 00:55:33,960 --> 00:55:37,520 on the left hand side where we want to kind of attract more 962 00:55:37,520 --> 00:55:40,360 security driven applications. And then like you know if that 963 00:55:40,360 --> 00:55:43,080 Shard is not there yet we we we might spare some effort in 964 00:55:43,080 --> 00:55:47,040 helping ECM to kind of build this solution for for for all 965 00:55:47,080 --> 00:55:48,520 for everyone in the whole ecosystem. 966 00:55:48,840 --> 00:55:51,720 So that's that's our goal like in the whole how we handle this 967 00:55:51,760 --> 00:55:55,760 situation and if if we really reach the point for where that 968 00:55:55,760 --> 00:55:58,720 Shard is not even enough. I do see like maybe some some 969 00:55:58,720 --> 00:56:02,280 other teams building layers 3 on top of us they can be validium 970 00:56:02,520 --> 00:56:05,680 they can kind of you know do other trade-offs that on top of 971 00:56:05,680 --> 00:56:10,440 us they can even use state diff host data on in a different way. 972 00:56:11,160 --> 00:56:13,760 But that that's what I see like you know how this world world 973 00:56:13,760 --> 00:56:17,240 goes it's like we will remain a very very high standard for 974 00:56:17,240 --> 00:56:21,120 security to attract the most kind of legit large the most 975 00:56:21,120 --> 00:56:24,520 fundamental applications. And then like if you if game 976 00:56:24,520 --> 00:56:28,440 really really require you know like high high throughput or 977 00:56:28,640 --> 00:56:32,280 like extremely high throughput or like that saying they can 978 00:56:32,360 --> 00:56:37,440 build a layer three as well idiom on top of us and like yeah 979 00:56:37,440 --> 00:56:39,760 I think it can be either can be layer three or some other other 980 00:56:39,760 --> 00:56:42,600 side solutions like at the same secure level. 981 00:56:42,600 --> 00:56:46,200 We were trying to kind of reduce the cost like when we are also 982 00:56:46,200 --> 00:56:48,680 working on data compression like how you kind of compress the 983 00:56:48,680 --> 00:56:51,760 data you post on chain. So there are some way like which 984 00:56:51,760 --> 00:56:55,200 which you can reduce your your data post on chain become like 985 00:56:55,200 --> 00:56:57,640 three times smaller. So there are some way for for 986 00:56:57,640 --> 00:56:59,720 doing compression. We are looking into that how you 987 00:56:59,720 --> 00:57:05,320 reduce the bridging cost but everything like stands like it 988 00:57:05,320 --> 00:57:07,360 can't. Like you know security should 989 00:57:07,600 --> 00:57:10,520 should always be the baseline and this would be never 990 00:57:10,520 --> 00:57:14,520 something we will compromise. Do do you think we'll be able to 991 00:57:14,520 --> 00:57:17,640 do data availability optimistically at some point? 992 00:57:18,160 --> 00:57:20,320 Because that would solve a lot of problems right. 993 00:57:21,480 --> 00:57:24,400 So I I I think it still depends on the like secure assumption I 994 00:57:24,400 --> 00:57:27,880 think for to me like you know I think for schools man chain 995 00:57:27,880 --> 00:57:30,200 like. So firstly I I don't see any 996 00:57:30,200 --> 00:57:33,720 kind of very very well established like optimistic 997 00:57:33,720 --> 00:57:36,600 database solution even if it's like you know there are teams 998 00:57:36,600 --> 00:57:38,840 like Igalias last year is building some DA solution. 999 00:57:39,360 --> 00:57:42,840 But I think it will still take some time to test whether this 1000 00:57:42,840 --> 00:57:45,840 works or not whether there's some other issue related to 1001 00:57:45,840 --> 00:57:48,320 because I know like you have been running this infrastructure 1002 00:57:48,320 --> 00:57:50,920 layer one you know like how many things subtle things that it 1003 00:57:50,920 --> 00:57:54,560 needed like there's Oracle RPC provider all those things might 1004 00:57:54,560 --> 00:57:58,520 require direct access to the data and ideally on the same 1005 00:57:58,520 --> 00:58:00,360 chain. So you don't know like if you 1006 00:58:00,360 --> 00:58:03,360 switch to the other solution what the subtle changes you need 1007 00:58:03,360 --> 00:58:07,480 to make and whether and I don't think they are all the most of 1008 00:58:07,480 --> 00:58:12,080 the solutions is mature that mature enough for for men 1009 00:58:12,400 --> 00:58:14,840 chained to migrate from. But I I think we are still 1010 00:58:14,840 --> 00:58:17,840 remaining optimistic. We will kind of always pay 1011 00:58:17,840 --> 00:58:20,960 attention to like the progress from other companies from from 1012 00:58:20,960 --> 00:58:24,200 other solutions. But right now we think you know 1013 00:58:24,200 --> 00:58:26,960 it's it's not there yet and we will stick to this kind of 1014 00:58:26,960 --> 00:58:29,800 security principle and encourage all the kind of layer three to 1015 00:58:29,800 --> 00:58:33,080 consider that as an option if you know you will have some 1016 00:58:33,120 --> 00:58:36,760 trust assumption. Yeah, you posited earlier that 1017 00:58:36,760 --> 00:58:41,920 there'll be a whole ecosystem of air tools with different trust 1018 00:58:41,920 --> 00:58:44,720 assumptions and different associated costs. 1019 00:58:45,040 --> 00:58:49,000 I agree that that's kind of like the the goal, but we're 1020 00:58:49,000 --> 00:58:50,960 currently very far away from that goal. 1021 00:58:51,000 --> 00:58:54,960 So just before we started recording this this podcast, a 1022 00:58:54,960 --> 00:58:58,000 tweet from Joseph Delong came out. 1023 00:58:58,400 --> 00:59:01,880 And it literally read L2's don't actually skate etherium, they 1024 00:59:01,880 --> 00:59:05,560 just fragmented into a bunch of unrelated chains obviously 1025 00:59:05,560 --> 00:59:16,200 alluding to the fact that into L2 bridges currently are not 1026 00:59:16,200 --> 00:59:18,520 operative. So basically it's a way to kind 1027 00:59:18,520 --> 00:59:21,960 of go between L2 is to kind of bridge or higher theorem, which 1028 00:59:21,960 --> 00:59:26,760 is obviously very costly. So in principle it seems like. 1029 00:59:27,680 --> 00:59:31,160 Bridging across different L2's should be possible because they 1030 00:59:31,160 --> 00:59:34,440 have the same security assumptions and kind of building 1031 00:59:34,440 --> 00:59:38,760 on top of Ethereum. When, when do you think we'll 1032 00:59:38,760 --> 00:59:40,560 get there and how do you see it happening? 1033 00:59:41,920 --> 00:59:45,240 Yeah that's a great question. So I think in terms of like in 1034 00:59:45,240 --> 00:59:49,360 turbo like between layer Two's, I do think like people are they 1035 00:59:49,360 --> 00:59:53,400 are like 2 directions like people are going towards it's 1036 00:59:53,400 --> 00:59:55,480 it's orthogonal it's not something that contradicted to 1037 00:59:55,480 --> 00:59:58,400 each other. SO11 approach is that as you 1038 00:59:58,400 --> 01:00:00,760 described like in all the layer tools are based on insurance so 1039 01:00:00,760 --> 01:00:03,080 they post the same state root on on ECM layer one. 1040 01:00:04,320 --> 01:00:07,240 So what you can do is that so for especially for the kill up 1041 01:00:07,360 --> 01:00:10,200 with with a faster like shorter finality that. 1042 01:00:11,360 --> 01:00:15,080 Imagine there's this exam layer, one you you post 22 state root 1043 01:00:15,080 --> 01:00:18,160 like the for for example there is scroll there is like the 1044 01:00:18,160 --> 01:00:22,760 other chain A&B and then like what you can do to access this 1045 01:00:22,760 --> 01:00:25,560 data is that you can read the state root of. 1046 01:00:26,240 --> 01:00:29,480 Of another layer 2 from ECM layer one because your route is 1047 01:00:29,480 --> 01:00:33,920 also there so you can read that and then provide a proof proving 1048 01:00:33,920 --> 01:00:37,240 that the the story slot the element you want to read from 1049 01:00:37,240 --> 01:00:40,880 the other layer two is a like provide this Merkel path to that 1050 01:00:40,880 --> 01:00:45,920 route and that you can trustlessly read and like 1051 01:00:45,960 --> 01:00:48,520 operate over the other chance like state. 1052 01:00:49,000 --> 01:00:51,840 So that's how you can do this. Yeah trust this way because 1053 01:00:52,040 --> 01:00:54,240 think about like you know two different layer ones because 1054 01:00:54,240 --> 01:00:56,760 they have totally different valid like validator set. 1055 01:00:56,920 --> 01:01:00,480 You you the bridge had to be kind of like multi seagull some 1056 01:01:00,480 --> 01:01:03,480 other like validator network where you don't have this kind 1057 01:01:03,480 --> 01:01:06,360 of shared security but for different layer tooth because 1058 01:01:06,360 --> 01:01:09,920 you're they could basically you can read the other chain state 1059 01:01:10,000 --> 01:01:13,000 through this proof like Merkel path from the other chain then 1060 01:01:13,000 --> 01:01:16,600 like but this seems to do some latency because you need some 1061 01:01:16,600 --> 01:01:20,520 time to generate proof and it might break the because because 1062 01:01:20,520 --> 01:01:25,320 of this this latency and so I do feel like in some sense it's 1063 01:01:25,320 --> 01:01:29,320 possible to have some standard in messaging between different 1064 01:01:29,320 --> 01:01:35,280 layer tools but it's still not super practical and user express 1065 01:01:35,280 --> 01:01:39,040 friendly to kind of do that here fully translated people still 1066 01:01:39,040 --> 01:01:43,840 use bridge to to bridge regardless another I I do feel 1067 01:01:43,840 --> 01:01:46,840 like you know as a proving technology has been improved as 1068 01:01:47,040 --> 01:01:50,080 as you know one layer to become adopting some similar standard 1069 01:01:50,080 --> 01:01:53,160 for how you how this communication can happen it 1070 01:01:53,160 --> 01:01:55,760 might be solved so that's one One Direction that direction 1071 01:01:55,760 --> 01:01:58,240 like people keep talking about is shared about shared sequencer 1072 01:01:58,560 --> 01:02:01,400 So shared sequencer is basically like where the same party 1073 01:02:01,400 --> 01:02:05,600 sequence to sequence like you know like transactions from two 1074 01:02:05,600 --> 01:02:08,800 different chance so that's how like that's where you can if you 1075 01:02:08,800 --> 01:02:12,120 want to interact with the other layer two and the same sequencer 1076 01:02:12,480 --> 01:02:17,920 know the order so that's like how you kind of maybe possible 1077 01:02:17,920 --> 01:02:20,360 like you have. This kind of you know some level 1078 01:02:20,360 --> 01:02:24,360 of compatibility between different chains but there are 1079 01:02:24,360 --> 01:02:27,360 like still a lot of problem related to this because mostly 1080 01:02:27,360 --> 01:02:29,720 shared sequencer are only ordering the transactions they 1081 01:02:29,720 --> 01:02:34,720 don't guarantee that like 1 transaction successfully done 1082 01:02:34,720 --> 01:02:37,160 like you execute the other it done has that level of 1083 01:02:37,680 --> 01:02:40,960 atomicity. So I like both are still very 1084 01:02:40,960 --> 01:02:42,880 early. I think we are we are kind of 1085 01:02:43,120 --> 01:02:46,360 watching closely to those those research directions but 1086 01:02:46,360 --> 01:02:48,320 currently it's just for example building building 1 chain. 1087 01:02:48,320 --> 01:02:50,640 I think eventually different layers will have different 1088 01:02:50,640 --> 01:02:54,000 communities and it's good to have some standard to talk with 1089 01:02:54,040 --> 01:02:57,440 each other because it's read at least compared with different 1090 01:02:57,440 --> 01:03:00,160 layer one fragmentation. It's theoretically possible to 1091 01:03:00,320 --> 01:03:05,480 to do that in a trustless way, but it it still takes some time. 1092 01:03:06,880 --> 01:03:09,280 I mean when you say different L twos will have different 1093 01:03:09,280 --> 01:03:13,160 communities, that's kind of at odds with what you said earlier, 1094 01:03:13,400 --> 01:03:15,680 that kind of different applications will choose 1095 01:03:15,680 --> 01:03:19,600 different L twos based on. The security assumptions they 1096 01:03:19,600 --> 01:03:22,240 need, right? So basically I may want to play 1097 01:03:22,800 --> 01:03:25,720 some game that doesn't need a lot of security assumptions, but 1098 01:03:25,720 --> 01:03:28,520 I still want my smart contract while it living on a different 1099 01:03:28,520 --> 01:03:30,360 chain, right? Yes, exactly. 1100 01:03:30,560 --> 01:03:33,040 Yes, Yes. I think due to this kind of 1101 01:03:33,040 --> 01:03:36,960 different property and assumptions like eventually 1102 01:03:36,960 --> 01:03:39,960 different layer tooth will have different like because they have 1103 01:03:39,960 --> 01:03:41,960 different value proposition. They have totally different 1104 01:03:41,960 --> 01:03:45,760 mission and vision like someone to kind of you know get users 1105 01:03:45,760 --> 01:03:49,120 through marketing and that may be part of their brand and some 1106 01:03:49,120 --> 01:03:51,720 might you know like want to build a totally different new 1107 01:03:51,720 --> 01:03:57,000 community from EVM and but we are like yeah we we we trying to 1108 01:03:57,000 --> 01:04:00,480 kind of be more like. But you don't see, you don't see 1109 01:04:00,480 --> 01:04:04,480 like the same users kind of using different chains for 1110 01:04:04,600 --> 01:04:08,720 different applications right now, yes, it's because most 1111 01:04:08,720 --> 01:04:11,880 chains don't have that many, like, you know, exciting 1112 01:04:11,880 --> 01:04:14,280 applications. It's like you know this same 1113 01:04:14,280 --> 01:04:19,720 suit for for defy and like but I do think like you know as you 1114 01:04:19,720 --> 01:04:21,880 know but but you you can feel like different chain had totally 1115 01:04:21,880 --> 01:04:25,240 different brand right Like some chance like the first impression 1116 01:04:25,240 --> 01:04:27,680 is this some some chance first impression. 1117 01:04:27,680 --> 01:04:31,200 This like first first impression for most people is like super 1118 01:04:31,200 --> 01:04:35,000 high SEM alignment, high very Regis for their tech and and and 1119 01:04:35,000 --> 01:04:38,000 and security assumptions. So if you go to kind of say if 1120 01:04:38,000 --> 01:04:41,840 you go to the field in Turkey or Malaysia or Nigeria or 1121 01:04:41,840 --> 01:04:44,960 Argentina, any of the countries you mentioned earlier, well 1122 01:04:44,960 --> 01:04:48,640 users who will actually use this to kind of solve real world 1123 01:04:48,880 --> 01:04:52,640 problems where they care which chain they're on. 1124 01:04:52,640 --> 01:04:55,840 So I in my view kind of the the DAP developer will kind of 1125 01:04:55,840 --> 01:04:59,720 decide what kind of is the best security or the needed security 1126 01:04:59,960 --> 01:05:03,200 model for their DAP and then the user ideally. 1127 01:05:03,760 --> 01:05:06,320 Down the line, they won't know which chain it's on, right? 1128 01:05:06,480 --> 01:05:08,040 They don't. They don't have to care about 1129 01:05:08,040 --> 01:05:10,680 the vibe. It's just like you and I browse 1130 01:05:10,680 --> 01:05:13,600 the Internet every day and we don't we don't really think 1131 01:05:13,600 --> 01:05:15,720 about TCPIP just happens in the background. 1132 01:05:16,240 --> 01:05:18,360 Yes, yes that's the most ideal situation. 1133 01:05:18,360 --> 01:05:20,840 So by developing communities there I'm more saying like in 1134 01:05:20,840 --> 01:05:23,840 the doing more like developer education and and growing like a 1135 01:05:23,840 --> 01:05:26,840 grassroot developer community because what what's different 1136 01:05:26,840 --> 01:05:29,240 like in developing community versus developing a global 1137 01:05:29,240 --> 01:05:33,320 community is there is that like local community can spot a lot 1138 01:05:33,320 --> 01:05:36,440 of like local problems because if I'm I'm I'm here I don't know 1139 01:05:36,440 --> 01:05:38,120 if I haven't traveled to those places. 1140 01:05:38,240 --> 01:05:40,280 I don't know like you know what people are building there like 1141 01:05:40,280 --> 01:05:42,000 what what problem people are facing. 1142 01:05:42,240 --> 01:05:46,280 So like getting developer there and like a track value of line 1143 01:05:46,400 --> 01:05:50,120 developer there grow community there can help getting more 1144 01:05:50,120 --> 01:05:52,600 local developer to build application on top of you and 1145 01:05:52,600 --> 01:05:54,440 then like potentially get more users. 1146 01:05:54,760 --> 01:05:57,800 So that's also like some. So that's how we grow. 1147 01:05:57,800 --> 01:06:00,480 This is not like we are acquiring users and then like 1148 01:06:00,480 --> 01:06:03,560 user have to choose their chain. But it's more like from 1149 01:06:03,680 --> 01:06:07,560 application level we can kind of help more grassroots developer 1150 01:06:07,560 --> 01:06:11,880 to to build to solve local problems and then like users 1151 01:06:11,880 --> 01:06:14,360 will use them and then use square as default chain. 1152 01:06:15,720 --> 01:06:18,080 So what does the road map look like for scroll? 1153 01:06:19,480 --> 01:06:22,320 I think the very very kind of short like technical road map 1154 01:06:22,320 --> 01:06:26,520 will be we will use data compression to massively reduce 1155 01:06:26,520 --> 01:06:30,280 the transaction cost. So like anyone can expect like 1156 01:06:30,400 --> 01:06:33,760 you know transactional scroll will become like you know way 1157 01:06:33,760 --> 01:06:37,920 cheaper like in in a few months. And why it takes a few months is 1158 01:06:37,920 --> 01:06:41,760 because we were trying to minimize the the frequency for 1159 01:06:41,760 --> 01:06:43,520 doing upgrade. Because doing upgrade is very 1160 01:06:44,040 --> 01:06:46,880 very kind of scary because you are using mostly things to 1161 01:06:46,880 --> 01:06:49,200 upgrade the most important part of your your protocol. 1162 01:06:49,880 --> 01:06:53,440 So being cheap using data compression is very important 1163 01:06:53,440 --> 01:06:56,560 part thing we will do. And second thing is that we are 1164 01:06:56,560 --> 01:06:58,520 keep focusing on improving the security. 1165 01:06:58,800 --> 01:07:01,760 So as you mentioned like censorship resistance, we are 1166 01:07:01,760 --> 01:07:05,960 trying to like you know have some some solution for that and 1167 01:07:05,960 --> 01:07:09,080 this kind of proposed A sequencer failure If if you know 1168 01:07:09,080 --> 01:07:11,960 we are not producing block then like what will happen. 1169 01:07:12,640 --> 01:07:14,960 So we are we are working on solutions for that that will 1170 01:07:14,960 --> 01:07:20,080 also like be solved like in in in around three months like you 1171 01:07:20,080 --> 01:07:23,240 know after that large upgrade what happened at that time and 1172 01:07:23,240 --> 01:07:26,880 then you will see like layer to beat which is a reference for 1173 01:07:26,880 --> 01:07:30,400 checking layer to security like a lot of red area will become 1174 01:07:30,400 --> 01:07:35,200 green And also we are exploring how to like even add more 1175 01:07:35,200 --> 01:07:37,320 security more than just what layer to be they're talking 1176 01:07:37,320 --> 01:07:40,640 about like people worry about like Ziki has bug, Ziki has bug. 1177 01:07:40,960 --> 01:07:45,520 So like we have exploring like multi multi prover system like 1178 01:07:45,520 --> 01:07:47,640 where we can add another additional prover. 1179 01:07:47,640 --> 01:07:51,040 So if they keep that work out the other prover can still be 1180 01:07:51,040 --> 01:07:54,160 back up to kind of like make sure like it's secure. 1181 01:07:54,160 --> 01:07:58,440 So security is very very important for us and then like 1182 01:07:58,440 --> 01:08:01,160 if application care about security they can come to us and 1183 01:08:01,160 --> 01:08:04,680 then that that will be a like short term goal. 1184 01:08:04,680 --> 01:08:07,880 And for the research we are we have been keep focusing on like 1185 01:08:07,880 --> 01:08:10,600 thisization, we will have some proposals published to kind of 1186 01:08:10,800 --> 01:08:15,360 discuss very broadly and we we are keep focusing on like 1187 01:08:15,360 --> 01:08:18,120 working on the next generation, the key VM to make the key and 1188 01:08:18,120 --> 01:08:20,840 become time time faster. There are already some we 1189 01:08:20,840 --> 01:08:23,399 designed there. We are still doing benchmark for 1190 01:08:23,760 --> 01:08:26,800 which process I want to use, how to architect the next generation 1191 01:08:26,800 --> 01:08:29,960 the KVM. So it's also there as a long 1192 01:08:29,960 --> 01:08:33,200 term, long term thing and in term of ecosystem community 1193 01:08:33,200 --> 01:08:35,399 building. I think we will get all the kind 1194 01:08:35,399 --> 01:08:38,359 of necessary infra as I mentioned ready like China just 1195 01:08:38,840 --> 01:08:42,640 integrated like days ago and the UNICO pass their governance and 1196 01:08:42,920 --> 01:08:45,920 a lot of like basic building blocks will be there like in a 1197 01:08:46,000 --> 01:08:49,200 few weeks or so like you know common build. 1198 01:08:49,640 --> 01:08:52,640 And then like we will have some like community initiative for 1199 01:08:52,880 --> 01:08:57,240 the community to to to interact to to kind of engage and more 1200 01:08:57,240 --> 01:09:01,920 'cause on on discord to kind of like really understand what what 1201 01:09:01,920 --> 01:09:05,319 our community really need. Yeah I I think a lot of 1202 01:09:05,319 --> 01:09:08,800 initiative will come out and also start community building in 1203 01:09:08,800 --> 01:09:12,920 the regions we we select and then like double down effort in 1204 01:09:13,240 --> 01:09:15,200 in kind of helping builders there to build. 1205 01:09:15,640 --> 01:09:19,399 So those are like a a short term like you know yeah and also like 1206 01:09:19,399 --> 01:09:21,399 internal security also definitely security console like 1207 01:09:21,399 --> 01:09:25,720 which we remove this what is it to some really credible security 1208 01:09:25,720 --> 01:09:29,240 console members that they will control like upgrade so that we 1209 01:09:29,240 --> 01:09:33,080 are not controlling like the the schools like upgrade. 1210 01:09:33,479 --> 01:09:36,439 So those are things where we're having short term like like 1211 01:09:36,439 --> 01:09:41,240 being cheaper ultra ultra like high security and also a lot of 1212 01:09:41,240 --> 01:09:43,760 community initiative and keep doing research. 1213 01:09:44,120 --> 01:09:48,720 So that's for that's what you can expect in in the next like 3 1214 01:09:48,720 --> 01:09:51,680 months. And I I, I know a lot of 1215 01:09:51,680 --> 01:09:53,600 projects are coming to to scroll ecosystem. 1216 01:09:53,600 --> 01:09:57,280 So like follow up like ecosystem account and scroll account and 1217 01:09:57,280 --> 01:09:59,760 scroll our our male facial account will man for protocol 1218 01:09:59,760 --> 01:10:03,520 upgrade update our our members event Hexas all those stuff 1219 01:10:03,720 --> 01:10:08,000 ecosystem we're talking about ecosystem project there's weekly 1220 01:10:08,000 --> 01:10:10,800 like what will happen. So if you know, feel free to 1221 01:10:10,800 --> 01:10:14,520 kind of interact and then like use what would like understand 1222 01:10:14,520 --> 01:10:17,640 what's happening there. And what's the best way to kind 1223 01:10:17,640 --> 01:10:20,640 of get in touch is, is there are all the links on your website. 1224 01:10:22,000 --> 01:10:25,000 The best way should be like our Twitter where Scroll underlines 1225 01:10:25,000 --> 01:10:28,400 AKP is our official account and then there's another ecosystem 1226 01:10:28,400 --> 01:10:31,120 account which is build with scroll build on scroll. 1227 01:10:31,120 --> 01:10:34,120 I need to check but that's that's like we have another 1228 01:10:34,120 --> 01:10:37,640 ecosystem account for following the recent ecosystem updates. 1229 01:10:37,960 --> 01:10:41,000 So if you follow that very closely, there's weekly updates 1230 01:10:41,000 --> 01:10:44,560 around everything happening on Scroll and we will support multi 1231 01:10:44,560 --> 01:10:47,680 language community. So if you are Chinese you are 1232 01:10:47,680 --> 01:10:51,080 like Spanish speaking Turkey speaking like it. 1233 01:10:51,960 --> 01:10:54,160 Like there will be like corresponding account like you 1234 01:10:54,160 --> 01:10:58,280 can follow to know more about school and join our discord. 1235 01:10:58,280 --> 01:11:00,480 Definitely yeah. Perfect. 1236 01:11:00,600 --> 01:11:02,280 Thank you so much for joining us today. 1237 01:11:04,640 --> 01:11:06,400 Thank you for joining us on this week's episode. 1238 01:11:06,760 --> 01:11:08,320 We release new episodes every week. 1239 01:11:09,080 --> 01:11:11,760 You can find and subscribe to the show on iTunes, Spotify, 1240 01:11:11,760 --> 01:11:14,800 YouTube, SoundCloud, or wherever you listen to podcasts. 1241 01:11:15,240 --> 01:11:18,040 And if you have a Google Home or Alexa device, you can tell it to 1242 01:11:18,040 --> 01:11:21,000 listen to the latest episode of the Epicenter podcast, go to 1243 01:11:21,000 --> 01:11:24,160 epicenter.tv/subscribe for a full list of places where you 1244 01:11:24,160 --> 01:11:26,400 can watch and listen. And while you're there, be sure 1245 01:11:26,400 --> 01:11:28,880 to sign up for the newsletter so you get new episodes in your 1246 01:11:28,880 --> 01:11:31,960 inbox as they're released. If you want to interact with us 1247 01:11:32,280 --> 01:11:34,680 guest or other podcast listeners, you can follow us on 1248 01:11:34,680 --> 01:11:36,840 Twitter and please leave us a review on iTunes. 1249 01:11:36,880 --> 01:11:38,960 It helps people find the show and we're always. 1250 01:11:39,080 --> 01:11:41,760 Happy to read them. So thanks so much and we look 1251 01:11:41,760 --> 01:11:42,920 forward to being back next week.