1 00:00:00,040 --> 00:00:04,560 This is Epicenter Episode 511 with guest Stefan Ghoslan. 2 00:00:18,920 --> 00:00:21,000 Welcome to Epicenter, the show which talks about the 3 00:00:21,000 --> 00:00:23,840 technologies, projects, and people driving decentralization 4 00:00:23,840 --> 00:00:27,360 and the boxing revolution. I'm Brian Frayn and I'm here 5 00:00:27,360 --> 00:00:29,750 with Phoenix Lutch. And today we're gonna speak with 6 00:00:29,750 --> 00:00:33,390 Stefan Ghosla. He is the founder of Frontier 7 00:00:33,390 --> 00:00:37,510 Research. Frontier is a sort of new me 8 00:00:37,510 --> 00:00:41,470 focused company, but he was also previously one of the Co 9 00:00:41,470 --> 00:00:45,310 founders of Flash Bots and I think sort of the lead architect 10 00:00:45,350 --> 00:00:47,790 I think if that's correct or product designers something like 11 00:00:47,790 --> 00:00:50,390 that at Flash Bots. So thanks so much for joining 12 00:00:50,390 --> 00:00:53,350 us, Stefan. Yeah, thank you. 13 00:00:53,750 --> 00:00:55,950 It's always a pleasure to get the chance to come back on this 14 00:00:55,950 --> 00:00:58,890 podcast and chat with you guys. Yeah, absolutely. 15 00:00:58,890 --> 00:01:00,930 It's not the first time, actually. 16 00:01:00,930 --> 00:01:03,170 Maybe you can just get started in there. 17 00:01:03,170 --> 00:01:08,130 I think it's a great question. So 2020, you wrote the blog post 18 00:01:08,130 --> 00:01:11,610 or a post on each research titled Flash Bots front running 19 00:01:11,610 --> 00:01:14,450 the MEV Crisis. So this is now three years ago 20 00:01:15,250 --> 00:01:19,370 and I'm curious just if you can like reflect a little bit sort 21 00:01:19,370 --> 00:01:22,810 of looking back. I mean you, you called it back 22 00:01:22,810 --> 00:01:27,340 then like MEV Crisis. I'm curious, You know what? 23 00:01:27,340 --> 00:01:30,700 What did you what was your understanding of, like, the Mev 24 00:01:30,780 --> 00:01:32,740 crisis? And like, what does it look like 25 00:01:32,740 --> 00:01:34,820 today? Are we in an Mev crisis? 26 00:01:36,660 --> 00:01:39,180 I was supposed to, like, think back on this, actually. 27 00:01:39,180 --> 00:01:42,540 Earlier today someone asked like, do we actually have a 28 00:01:42,540 --> 00:01:45,140 counterfactual to PBS? Like, what would have happened 29 00:01:45,140 --> 00:01:47,460 if we hadn't launched Math Boost? 30 00:01:47,740 --> 00:01:50,220 Would we actually have a significant different 31 00:01:50,220 --> 00:01:54,380 distribution of concentration on the validator side? 32 00:01:55,140 --> 00:01:56,980 Or would it look like pretty similar? 33 00:01:57,660 --> 00:02:04,980 The MEV crisis was basically the thesis that said if there wasn't 34 00:02:05,580 --> 00:02:12,460 action taken by the community and early Flash bots members who 35 00:02:12,540 --> 00:02:16,660 sort of felt responsible to take this this action, there would be 36 00:02:16,940 --> 00:02:20,460 an increase in concentration of hashing power because. 37 00:02:21,660 --> 00:02:25,740 The level of activity on Ethereum was increasing to the 38 00:02:25,740 --> 00:02:27,460 point where MEV became meaningful. 39 00:02:28,140 --> 00:02:34,420 I if you remember during summer 2020 defy summer we had all of 40 00:02:34,420 --> 00:02:38,700 the sort of start of the yield farming and like really you to 41 00:02:38,700 --> 00:02:41,380 swap getting traction for the first time with all the 42 00:02:41,380 --> 00:02:46,820 liquidity mining programs is it yam finance and and all this 43 00:02:46,860 --> 00:02:51,800 this this good stuff. And so it was clear that, you 44 00:02:51,800 --> 00:02:56,360 know, whether it's token sniping, arbitrage, whatever 45 00:02:56,360 --> 00:02:58,560 else. There was an advantage of 46 00:02:58,600 --> 00:03:01,920 operating this at the validator level and validators weren't 47 00:03:01,920 --> 00:03:04,280 currently doing it. The activity was happening 48 00:03:04,280 --> 00:03:07,960 through the transaction pool, through sort of gaming the 49 00:03:08,680 --> 00:03:11,680 algorithm that the miners run for packing blocks. 50 00:03:13,200 --> 00:03:16,720 So the crisis was, well, it's very clear that miners have the 51 00:03:16,720 --> 00:03:20,610 advantage in doing these games. They are likely to start to 52 00:03:20,610 --> 00:03:23,690 invest into running things themselves or run custom 53 00:03:23,690 --> 00:03:26,490 clients, make partnerships with trading firms to be able to 54 00:03:26,490 --> 00:03:29,770 extract some of that value. And that has the potential to 55 00:03:29,770 --> 00:03:32,730 lead towards more concentration because there's sort of 56 00:03:32,730 --> 00:03:35,570 increasing returns to being the ones that play this role. 57 00:03:36,370 --> 00:03:39,490 Maybe there is a better mechanism that we can develop 58 00:03:39,490 --> 00:03:44,090 and launch that enables to maintain transparency? 59 00:03:45,020 --> 00:03:50,980 Over what is sort of happening in the system and maintain Open 60 00:03:50,980 --> 00:03:54,260 Access to it. So make it so it's not doesn't 61 00:03:54,260 --> 00:03:56,340 require a special deal with validators, but you can freely 62 00:03:56,340 --> 00:03:59,700 connect and participate. And finally have some way to 63 00:03:59,700 --> 00:04:03,260 sort of redistribute the value to the actors in a way that's 64 00:04:03,260 --> 00:04:06,220 like welfare maximizing or whatever definition that you can 65 00:04:06,220 --> 00:04:08,650 have. So that's the Mev crisis. 66 00:04:09,290 --> 00:04:11,770 You sort of mentioned like, hey, you wrote this post and you made 67 00:04:11,770 --> 00:04:14,050 it on E3 search. And then I'm like thinking back 68 00:04:14,050 --> 00:04:16,730 to the last three years of my life and I feel like all I did 69 00:04:16,730 --> 00:04:20,890 was write 2 posts and like everything else was like fluff, 70 00:04:21,089 --> 00:04:22,770 right? Like I wrote this front running 71 00:04:22,770 --> 00:04:27,010 the Mev crisis, which was like a spec for Mevgath, which was like 72 00:04:27,010 --> 00:04:30,130 the client that that we developed and then ended up 73 00:04:30,130 --> 00:04:34,310 getting adopted by minors. And the second one that I wrote 74 00:04:34,310 --> 00:04:40,550 was introducing Mev Boost as a as a spec and then maybe like I 75 00:04:40,550 --> 00:04:42,910 should have just written those things and just like you know 76 00:04:42,910 --> 00:04:45,190 went to the beach and did nothing else the rest of the 77 00:04:45,230 --> 00:04:47,790 time. Because like those specs alone I 78 00:04:47,790 --> 00:04:50,590 feel like move the needle into like what the infrastructure 79 00:04:50,590 --> 00:04:54,950 looks like and and and we're like the most impactful things 80 00:04:54,950 --> 00:04:59,830 that that I really did. So yeah, I sort of there's this 81 00:04:59,830 --> 00:05:01,830 like, really fun progression from, you know? 82 00:05:02,230 --> 00:05:04,870 There not being any infrastructure to now having 83 00:05:04,870 --> 00:05:09,150 gone through two major cycles of the Mev infrastructure changing 84 00:05:09,150 --> 00:05:14,230 on Ethereum and do you feel like these both of these things had 85 00:05:14,230 --> 00:05:16,310 the impact that you hoped it would? 86 00:05:17,150 --> 00:05:20,470 In terms of mitigating the Mev crisis, yes, yeah. 87 00:05:22,830 --> 00:05:28,380 Not really. Not really. 88 00:05:28,380 --> 00:05:30,940 I mean it's it's like hard to say like there's no way else. 89 00:05:30,940 --> 00:05:34,620 Is that what's happened or? Yeah. 90 00:05:34,620 --> 00:05:38,020 I mean it's like it achieves the like the outcome in terms of 91 00:05:38,220 --> 00:05:40,700 like it's a system like I adopted and like everyone got 92 00:05:40,700 --> 00:05:45,420 excited about it and and you know introduce a new industry 93 00:05:45,420 --> 00:05:48,100 and ecosystem with a lot of people that are having open 94 00:05:48,100 --> 00:05:50,740 discussions about these things and implementing alternative 95 00:05:51,100 --> 00:05:53,900 solutions to them. That may be crisis in itself. 96 00:05:53,900 --> 00:05:56,980 I mean, one can debate whether it's a crisis at all, if it ever 97 00:05:56,980 --> 00:06:00,780 was actually there, or if it's just like a natural way that an 98 00:06:00,780 --> 00:06:04,060 industry evolves and whether, you know. 99 00:06:04,510 --> 00:06:07,430 Whether it was Metgath, met Boost or something else would 100 00:06:07,430 --> 00:06:10,110 have happened, the end results would have been the same and 101 00:06:10,110 --> 00:06:11,830 we've ended up in the same location. 102 00:06:12,270 --> 00:06:14,910 And at the at the end of the day, I don't know that the 103 00:06:14,910 --> 00:06:18,350 amount of concentration that we have at the stake level today is 104 00:06:18,350 --> 00:06:21,270 like meaningfully different than what we would have if we didn't 105 00:06:21,270 --> 00:06:23,910 have PBS or like any sort of MDB infrastructure. 106 00:06:24,580 --> 00:06:28,140 Maybe we have more transparency in how the system works because 107 00:06:28,140 --> 00:06:31,260 you know, we have sort of this open pricing mechanism where the 108 00:06:31,260 --> 00:06:33,660 information is being like routed through. 109 00:06:33,660 --> 00:06:36,460 We have like a dollar value for like the price of blocks. 110 00:06:36,460 --> 00:06:38,500 And you have some nice dashboards that like allow us to 111 00:06:38,500 --> 00:06:41,180 show it like maybe there was some path dependency there 112 00:06:41,180 --> 00:06:43,860 that's different. But in terms of like how things 113 00:06:43,860 --> 00:06:47,340 are evolving, it seems like pretty inherent to the way that 114 00:06:47,620 --> 00:06:51,740 blockchains are designed that the end result will end up being 115 00:06:52,300 --> 00:06:55,100 being very similar. Maybe maybe just maybe just 116 00:06:55,100 --> 00:07:01,180 commenting here one thing, I at least I do feel like reasonably 117 00:07:01,940 --> 00:07:06,980 confident that MEV did not increase validator 118 00:07:07,500 --> 00:07:11,900 centralization in a meaningful like was not a meaningful driver 119 00:07:11,900 --> 00:07:15,060 for validator centralization on this area. 120 00:07:15,140 --> 00:07:17,620 I think that's pretty clear. Yeah. 121 00:07:18,940 --> 00:07:21,460 So I think from that perspective, if you take, if you 122 00:07:21,460 --> 00:07:27,420 take like concentration on the validator level, then I think if 123 00:07:27,420 --> 00:07:29,740 that's the kind of criteria, then I would say like, yeah, I 124 00:07:29,740 --> 00:07:34,860 think it worked out, right, like that didn't happen or it 125 00:07:34,860 --> 00:07:38,020 happened to some extent for other reasons that were 126 00:07:38,020 --> 00:07:40,780 unrelated to MUV, I think. Right, right. 127 00:07:41,100 --> 00:07:45,460 It didn't fail, yeah. We don't know if it was 128 00:07:45,460 --> 00:07:48,020 necessary, but we know that it didn't fail. 129 00:07:49,310 --> 00:07:51,510 I think the thing that I'm the most excited about in the end 130 00:07:51,630 --> 00:07:56,670 and in particular with the introduction of of Mevboost like 131 00:07:56,670 --> 00:07:59,990 so Mevgeth was like a really simple solution. 132 00:07:59,990 --> 00:08:05,430 It was like get Flashboss to run a server and then like route all 133 00:08:05,430 --> 00:08:09,270 the Mevs through the server and then like Multiplex it to the 134 00:08:09,270 --> 00:08:12,550 minors, right? And it was just like introduce 135 00:08:12,550 --> 00:08:16,830 like Flashboss has like a server that like runs the Mev market. 136 00:08:17,360 --> 00:08:23,120 Mevboost is a much more sort of open system where you have a lot 137 00:08:23,120 --> 00:08:27,600 of different layers to it and a lot of room for individual 138 00:08:27,600 --> 00:08:30,920 entities to compete and offering additional services, right. 139 00:08:30,920 --> 00:08:34,480 So the just the introduction of the the block builder role and 140 00:08:34,600 --> 00:08:40,840 the relay role has sort of meant that this thing of operating an 141 00:08:40,840 --> 00:08:43,640 MEV market that the Flashbluts was doing now is being done by 142 00:08:43,640 --> 00:08:46,200 like I don't know 6-7 eight companies. 143 00:08:46,900 --> 00:08:51,980 And there's quite a bit of turnover I think in the like in 144 00:08:51,980 --> 00:08:54,980 the dominance of of these parties. 145 00:08:55,300 --> 00:08:58,500 They're like trying to figure out how to like make it 146 00:08:58,500 --> 00:09:02,140 sustainable, how to have a revenue model and competing for 147 00:09:02,140 --> 00:09:06,020 features that they develop. I think that's all very positive 148 00:09:06,020 --> 00:09:09,700 because it brings a lot more diversity to the way that these 149 00:09:09,700 --> 00:09:15,070 systems are being iterated upon. And just the number of different 150 00:09:15,070 --> 00:09:17,830 perspectives they're being reflected into the the 151 00:09:17,830 --> 00:09:19,310 architectures that they get adopted. 152 00:09:19,790 --> 00:09:22,790 So that that to me has been like quite successful, if nothing 153 00:09:22,790 --> 00:09:25,790 else, just having more different stakeholders are involved in 154 00:09:25,790 --> 00:09:29,350 developing and competing against each other for developing this 155 00:09:29,630 --> 00:09:32,510 this infrastructure, providing services on top of it, that 156 00:09:32,510 --> 00:09:36,470 seems like quite a decentralized way to approach infrastructure 157 00:09:36,470 --> 00:09:39,470 development. So it's almost like the third 158 00:09:39,470 --> 00:09:42,070 thing you did right? Like I guess formalizing this. 159 00:09:42,580 --> 00:09:45,460 Mev supply chain with the different actors and how you can 160 00:09:46,100 --> 00:09:49,500 break them apart and then how you can like separate the roles. 161 00:09:50,500 --> 00:09:54,220 Did anything change in that perspective from how you looked 162 00:09:54,220 --> 00:09:58,340 at the Mev supply chain during Mev Boost and like nowadays? 163 00:09:59,060 --> 00:10:01,420 Or has anything you haven't foreseen or something I. 164 00:10:03,860 --> 00:10:06,300 Don't think there's anything that I haven't foreseen. 165 00:10:06,780 --> 00:10:10,280 I think what? I've been trying to introduce as 166 00:10:10,280 --> 00:10:13,600 like a reframing of the Mev supply chain is to think of it 167 00:10:13,600 --> 00:10:17,280 less linearly and think of it more as sort of a chaotic set of 168 00:10:17,280 --> 00:10:19,560 games. I presented a little while ago 169 00:10:19,560 --> 00:10:25,240 and we wrote this blog post with Frontier about infinite games 170 00:10:25,400 --> 00:10:29,840 and introducing this idea of a transaction supply network. 171 00:10:30,360 --> 00:10:33,960 So instead of thinking of Mev supply chain as being like users 172 00:10:33,960 --> 00:10:38,320 originated MEV and then. It sort of gets sort of funneled 173 00:10:38,320 --> 00:10:40,840 through this system of of extraction all the way to the 174 00:10:40,840 --> 00:10:45,880 validator inside be inside. Do you think about a cluster of 175 00:10:45,960 --> 00:10:47,800 a bunch of different infrastructure component that 176 00:10:47,800 --> 00:10:51,880 yeah, it originates with with users, but depending on gets 177 00:10:51,880 --> 00:10:55,000 either intense like whatever they're trying to achieve and 178 00:10:55,040 --> 00:10:58,280 interacting with the chain, they get routed through different 179 00:10:58,280 --> 00:11:01,320 sets of specialized infrastructure that leverage 180 00:11:01,320 --> 00:11:04,480 tools like auctions. I leverage tools like privacy. 181 00:11:05,680 --> 00:11:10,080 To provide those services and capture that those system that 182 00:11:11,280 --> 00:11:14,480 reflect that value back to the user and provide the utility 183 00:11:14,480 --> 00:11:17,360 that the user is looking for. It's kind of like this, this 184 00:11:17,520 --> 00:11:20,040 thing to say, like, OK, MEV is like a lens through which you 185 00:11:20,040 --> 00:11:23,680 can see the world, right? And once you start to look at 186 00:11:23,680 --> 00:11:28,160 like bridges, you look at changes, you look at all these 187 00:11:28,160 --> 00:11:29,640 different things from the Mev lens. 188 00:11:29,640 --> 00:11:32,360 It's like a toolkit that allows you to analyze them and 189 00:11:32,360 --> 00:11:36,450 understand how. How do the dynamics play out if 190 00:11:36,930 --> 00:11:39,130 you set this up in a decentralized setting? 191 00:11:40,450 --> 00:11:44,730 I think it's useful to have this lens, but still look at all 192 00:11:44,730 --> 00:11:48,770 these components individually, all these games individually, as 193 00:11:48,770 --> 00:11:52,330 opposed to say, there's only one game here and it's like the Mev 194 00:11:52,330 --> 00:11:54,810 supply chain and that's the only thing that matters. 195 00:11:55,310 --> 00:11:57,670 I'd rather look at individual components. 196 00:11:57,870 --> 00:12:00,910 Can you talk a bit about your vision for frontier research? 197 00:12:00,910 --> 00:12:04,710 Like what do you hope to build? What kind of organization do you 198 00:12:04,710 --> 00:12:09,550 hope to build, and what kind of impact do you hope to have on 199 00:12:10,590 --> 00:12:13,350 the ME landscape or crypto more generally? 200 00:12:15,190 --> 00:12:19,950 Yeah, Frontier Research. So I left flash bots at the end 201 00:12:19,950 --> 00:12:28,330 of. 2022 like and beginning of the fall 2022 and I so took some 202 00:12:28,330 --> 00:12:30,370 time off and decided like is this something I want to 203 00:12:30,370 --> 00:12:33,930 continue working on, do I want to continue working on on MUVI 204 00:12:33,930 --> 00:12:36,530 felt like that there was still some contributions I had to make 205 00:12:36,530 --> 00:12:41,290 to the space and so started sort of a research organization 206 00:12:41,290 --> 00:12:44,450 around this that were helps working with various different 207 00:12:44,450 --> 00:12:48,370 teams that are that are participating across the entire. 208 00:12:49,350 --> 00:12:53,590 The entire ecosystem of MVV infrastructure help them sort of 209 00:12:53,590 --> 00:12:57,470 up level the level, the their understanding and mental models 210 00:12:57,590 --> 00:13:01,590 around MV help do some analysis over the system that they 211 00:13:01,590 --> 00:13:04,070 produce and advise on go to market and whatnot. 212 00:13:04,910 --> 00:13:08,390 So we've been doing that for quite a few months now since 213 00:13:08,790 --> 00:13:13,510 since the beginning of of this year now we've we've started 214 00:13:13,510 --> 00:13:16,550 looking more active like OK, where are the opportunities for 215 00:13:16,550 --> 00:13:20,460 developing products. We've started incubating a block 216 00:13:20,460 --> 00:13:25,500 builder called called Faith Builder that participates in 217 00:13:25,500 --> 00:13:30,060 this game. I sort of see very abstractly 218 00:13:30,860 --> 00:13:35,580 the set of games as being an abstract set of two two roles. 219 00:13:35,900 --> 00:13:39,460 You either have a message passing system with some rules 220 00:13:39,620 --> 00:13:43,820 over how these messages are handled and aggregated and you 221 00:13:43,820 --> 00:13:48,500 have individual agents. That can be called like solvers 222 00:13:48,700 --> 00:13:53,020 or searchers or block builders or whatever else that aggregate 223 00:13:53,020 --> 00:13:56,260 these messages and then solve them, optimize them according to 224 00:13:56,260 --> 00:14:00,340 some objective function. And so the block builder that 225 00:14:00,780 --> 00:14:04,300 we're running we see as being sort of fundamental to enabling 226 00:14:04,540 --> 00:14:06,660 the development of all these different games across the 227 00:14:06,660 --> 00:14:09,340 ecosystem and participating in them and helping them bootstrap 228 00:14:10,100 --> 00:14:12,100 to provide the better user services. 229 00:14:12,620 --> 00:14:16,270 So it's. The generalized solver that aims 230 00:14:16,270 --> 00:14:21,190 to participate in all of these different blockchain games. 231 00:14:22,390 --> 00:14:27,350 So yeah, that's sort of a venture that we're incubating 232 00:14:27,350 --> 00:14:32,670 from the Frontier team. So this means like it's a 233 00:14:32,670 --> 00:14:35,670 builder in the flash was muv boost because somebody could 234 00:14:35,670 --> 00:14:40,270 also be a solver maybe know, I don't know for cow swap or like 235 00:14:40,270 --> 00:14:42,150 you know other products like that. 236 00:14:42,150 --> 00:14:45,970 Is that correct? There are, they're very similar 237 00:14:45,970 --> 00:14:50,810 games at the end of the day, they they achieve similar 238 00:14:50,810 --> 00:14:55,250 objectives and they sort of are architected and in ways that 239 00:14:55,250 --> 00:14:58,570 benefit from participation in multiple of these of these 240 00:14:58,570 --> 00:15:02,850 games. I mean, the move of activity and 241 00:15:02,850 --> 00:15:05,210 like the trend, this is like 1 trend that's that's interesting 242 00:15:05,210 --> 00:15:10,690 to discuss is how, how much Rfqs are going to gain in momentum 243 00:15:10,690 --> 00:15:15,110 and adoption. Over the course of the next 12 244 00:15:15,110 --> 00:15:20,350 months, can you explain Rfqs? Yeah, I can give like. 245 00:15:21,150 --> 00:15:23,310 Maybe I can give like an mev explanation to it. 246 00:15:24,110 --> 00:15:25,110 Sure. Please, Yeah. 247 00:15:25,910 --> 00:15:30,070 And then I'll give the like the maybe I'll give the user 248 00:15:30,070 --> 00:15:31,910 explanation and then the MEV explanation as well. 249 00:15:31,910 --> 00:15:37,070 So the user explanation is you're used to doing trades on 250 00:15:37,070 --> 00:15:41,510 top of D5 by going to usersoft.com and like requesting 251 00:15:41,510 --> 00:15:43,440 a swap. What happens then is you're 252 00:15:43,440 --> 00:15:49,160 tapping onto AMM liquidity that's on chain, but you have a 253 00:15:49,160 --> 00:15:52,880 lot of inefficiency in the prices that you obtained. 254 00:15:53,200 --> 00:15:58,560 So the liquidity that's on chain is stale, it isn't by 255 00:15:58,560 --> 00:16:02,920 professional counterparties and you're susceptible to 256 00:16:04,240 --> 00:16:07,240 sanctioning attacks and like exploitation of your orders if 257 00:16:07,240 --> 00:16:11,400 you if you, if you route it through there, so. 258 00:16:12,240 --> 00:16:16,240 It doesn't give you very good pricing essentially to interact 259 00:16:16,240 --> 00:16:22,440 with uni swap relative to the pricing that professional 260 00:16:22,440 --> 00:16:25,240 traders would be able to get by interacting with all the 261 00:16:25,240 --> 00:16:29,920 different liquidity venues. RFQ just allows you to outsource 262 00:16:30,080 --> 00:16:34,040 the work of routing through a bunch of different liquidity 263 00:16:34,040 --> 00:16:39,320 venues to some solving system it says hey, we're going to 264 00:16:39,320 --> 00:16:41,700 introduce. A marketplace of service 265 00:16:41,700 --> 00:16:43,900 providers that are going to compete to give you the best 266 00:16:43,900 --> 00:16:48,860 price, and they can do the complex work of figuring out how 267 00:16:48,860 --> 00:16:53,420 to access liquidity wherever it is on the user side. 268 00:16:53,420 --> 00:16:57,660 It's great you have the same experience of doing a swap and 269 00:16:57,660 --> 00:17:02,700 getting the tokens that you want, but you can also expect to 270 00:17:02,700 --> 00:17:07,020 get better pricing on the trades that you do. 271 00:17:07,780 --> 00:17:12,780 The Mev lens is. Rfq's like recapture value 272 00:17:12,900 --> 00:17:15,980 that's being leaked sort of down the supply chain, right. 273 00:17:16,900 --> 00:17:20,140 So if you look at the distribution of validator 274 00:17:20,140 --> 00:17:26,020 revenues today, over 50% of it comes from basically arbitrage 275 00:17:26,140 --> 00:17:31,620 against unit swap Lp's, right. And so US swap Lp's put a bunch 276 00:17:31,620 --> 00:17:34,900 of capital into into their AMM pools. 277 00:17:35,350 --> 00:17:39,990 And they're relying on C Phi, D Phi arbitragers to update the 278 00:17:39,990 --> 00:17:44,990 prices that they offer relative to the prices on other liquidity 279 00:17:44,990 --> 00:17:46,630 venues like by an Anson and whatnot. 280 00:17:47,830 --> 00:17:50,670 And arbitragers are capturing significant amount of value. 281 00:17:51,110 --> 00:17:56,310 The Ammlps are losing that value and this is called LVR loss 282 00:17:56,310 --> 00:17:59,310 versus rebalancing. There's a service that these 283 00:17:59,310 --> 00:18:02,550 arbitragers are offering which is to maintain prices on chain. 284 00:18:02,870 --> 00:18:04,950 It comes at a significant cost for. 285 00:18:05,750 --> 00:18:09,230 The Lps and a lot of that value ends up being captured by 286 00:18:09,630 --> 00:18:13,390 validators. What RFQ does is it says, OK, 287 00:18:13,390 --> 00:18:16,270 thank you very much for like the system that was like generating 288 00:18:16,270 --> 00:18:20,510 all our value for validators. Like we as the originators of 289 00:18:20,510 --> 00:18:23,550 this order flow would rather recapture this value either for 290 00:18:23,550 --> 00:18:26,390 ourselves or for our users. And so we're going to create the 291 00:18:26,390 --> 00:18:29,710 system that enables us to internalize essentially the 292 00:18:29,710 --> 00:18:34,230 liquidity. And provide that value back to 293 00:18:34,630 --> 00:18:38,990 to our users. So the sort of trend on on this 294 00:18:38,990 --> 00:18:43,390 RFQ stuff is for more and more of the liquidity to move off 295 00:18:43,390 --> 00:18:47,590 chain, to move away from from AMM contracts and be sort of 296 00:18:47,590 --> 00:18:50,470 accessible by these professional market makers, these 297 00:18:50,470 --> 00:18:54,870 professional solvers that can get better price improvements to 298 00:18:54,870 --> 00:18:57,030 the users and better quotes to the users. 299 00:18:58,270 --> 00:19:03,190 And and expose less of that, sort of down towards towards 300 00:19:03,270 --> 00:19:07,750 others. And then in this RFQ example, so 301 00:19:07,750 --> 00:19:11,990 let's say I'm on Ethereum, you know I wanna trade some E for 302 00:19:11,990 --> 00:19:17,070 USCC for example. Then I'm making this request for 303 00:19:17,510 --> 00:19:23,310 quotation or something and then it's like hey, I want I have E 304 00:19:23,310 --> 00:19:28,550 if I want USCC it it this also means that somebody could come 305 00:19:28,550 --> 00:19:33,910 and say, hey I'm a trading company and I can make that 306 00:19:33,910 --> 00:19:39,990 trade on finance and here's how many, you know USCC I will give 307 00:19:39,990 --> 00:19:43,510 you for this EVE and you know it nothing ever settle. 308 00:19:43,550 --> 00:19:47,630 I mean of course they're going to send me the USCC on Ethereum, 309 00:19:48,150 --> 00:19:51,910 but but the trade actually gets executed on finance. 310 00:19:51,910 --> 00:19:55,640 So this would be this will be possible or? 311 00:19:57,040 --> 00:20:00,160 That's like, yeah, that's mostly how Rfqs work. 312 00:20:01,360 --> 00:20:04,000 Traders don't necessarily need to make the trade on finance at 313 00:20:04,000 --> 00:20:06,960 the end of the day, like they hold a bunch of tokens on chain 314 00:20:07,360 --> 00:20:11,120 and they like heads their exposure on other venues. 315 00:20:12,640 --> 00:20:16,640 So basically it's like they have inventory on Ethereum that they 316 00:20:16,640 --> 00:20:20,560 are providing within their market making and their goal is 317 00:20:20,560 --> 00:20:22,560 to keep this inventory within some range. 318 00:20:23,100 --> 00:20:26,620 Of parameters that's like in line with their hedging on the 319 00:20:26,620 --> 00:20:28,820 centralized exchange. And yeah, whenever this 320 00:20:28,820 --> 00:20:31,340 inventory goes like too far in One Direction or another that 321 00:20:31,340 --> 00:20:33,860 they lose money, but so long as it stays within a certain range 322 00:20:33,860 --> 00:20:37,660 that they make line. Maybe let's talk about block 323 00:20:37,660 --> 00:20:41,060 building. I feel like block building is 324 00:20:41,060 --> 00:20:45,140 something that's. Very poorly understood I think 325 00:20:45,140 --> 00:20:50,380 by most people even it probably even E theorem uses ever like no 326 00:20:50,380 --> 00:20:54,500 a year right that like valid or not actually building the blocks 327 00:20:54,500 --> 00:20:58,420 and happens in some other way and and and how that's 328 00:20:58,420 --> 00:21:01,340 happening. So can you just explain what 329 00:21:01,340 --> 00:21:05,700 what is the the what's block building and what is the the 330 00:21:05,780 --> 00:21:08,140 block building landscape in Ethereum at the moment. 331 00:21:09,700 --> 00:21:11,700 Yeah. So block building is is 332 00:21:11,700 --> 00:21:17,210 basically validators outsourcing the job of putting transactions 333 00:21:17,210 --> 00:21:22,250 into a block to a marketplace. Anyone can be a block builder. 334 00:21:22,250 --> 00:21:25,490 Anyone can like spin one up not everyone can be a good block 335 00:21:25,490 --> 00:21:29,370 builder because block builder is very dependent on having access 336 00:21:29,370 --> 00:21:33,010 to order flow. So there's a trend here, which 337 00:21:33,010 --> 00:21:37,330 is increasing amounts of private order flow that's being routed 338 00:21:37,330 --> 00:21:39,810 to block builders. Interestingly enough, right, 339 00:21:39,810 --> 00:21:42,370 like a lot of the discussion there was initially, like, oh, 340 00:21:42,370 --> 00:21:46,010 no, block builders are going to get exclusive order flow deals 341 00:21:46,410 --> 00:21:49,850 and become super dominant and it'll be centralizing towards 342 00:21:50,370 --> 00:21:52,370 towards the chain because you'll have just like one party that 343 00:21:52,370 --> 00:21:56,450 produces all the blocks. Realistically, I think that's 344 00:21:56,570 --> 00:21:59,450 there's a lot of economic dynamics that mean that that's 345 00:21:59,450 --> 00:22:03,250 not going to be the case. What is happening is there's an 346 00:22:03,250 --> 00:22:07,970 increasing adoption of order flow aggregation systems. 347 00:22:09,050 --> 00:22:12,210 So more and more of the transaction flows being routed 348 00:22:12,210 --> 00:22:17,330 through systems that say we will directly send to block builders, 349 00:22:17,330 --> 00:22:19,130 we'll send to all the block builders, but we'll directly 350 00:22:19,130 --> 00:22:22,290 send to the block builders and never touch the public 351 00:22:22,290 --> 00:22:24,690 transaction pool. And so essentially, if you're a 352 00:22:24,690 --> 00:22:27,410 block builder, you have to connect all these different 353 00:22:27,410 --> 00:22:31,290 private sources of order flow to be able to build competitive 354 00:22:31,290 --> 00:22:34,210 blocks. You compete on this access of 355 00:22:34,210 --> 00:22:37,390 the order flow, compete on the efficiency of the merging and 356 00:22:37,390 --> 00:22:42,990 then submit blocks to validators for inclusion. 357 00:22:43,470 --> 00:22:46,470 There's sort of two types of block builders. 358 00:22:46,750 --> 00:22:50,030 There's block builders that generate the order flow 359 00:22:50,030 --> 00:22:53,510 themselves, which you could call integrated sort of searcher 360 00:22:53,510 --> 00:22:56,710 builders. A lot of the dominant block 361 00:22:56,710 --> 00:22:59,790 builders to date have been of this kind. 362 00:23:00,990 --> 00:23:04,230 They don't really monetize from playing the role of the block 363 00:23:04,230 --> 00:23:06,630 builder. The block builder is sort of a 364 00:23:06,630 --> 00:23:11,150 way for them to execute trades that they would want to execute 365 00:23:11,150 --> 00:23:14,870 anyways and have some edge in providing this execution. 366 00:23:16,150 --> 00:23:18,510 And then the second type of block builders are sort of 367 00:23:18,510 --> 00:23:21,070 neutral block builders and they're the ones that don't 368 00:23:21,070 --> 00:23:24,350 generate order for themselves. They don't do trading 369 00:23:24,350 --> 00:23:27,550 themselves. They just aggregate the trading 370 00:23:27,550 --> 00:23:32,670 as done by a bunch of other searchers, traders and provide 371 00:23:32,910 --> 00:23:36,550 sort of Fair execution across all that flow. 372 00:23:37,230 --> 00:23:39,030 And that's like one of the big questions right now on the block 373 00:23:39,030 --> 00:23:43,430 building side is, is it inevitable for block builders to 374 00:23:43,510 --> 00:23:48,870 be traders or can neutral block builders provide services and be 375 00:23:48,870 --> 00:23:52,630 competitive? So if you think about a business 376 00:23:52,630 --> 00:23:56,230 model of the builder searcher, I guess they make the money 377 00:23:56,270 --> 00:24:00,030 through their own searching and the neutral block builder has to 378 00:24:00,030 --> 00:24:04,240 sort of. Come up with some way to to make 379 00:24:04,240 --> 00:24:06,440 money from the private order flow or how do you kind of 380 00:24:06,920 --> 00:24:09,280 attract the order flow just by being a neutral party? 381 00:24:09,280 --> 00:24:13,600 Is that basically the the strategy or you know I guess 382 00:24:13,600 --> 00:24:16,160 whatever you can share or how how do they compete with this 383 00:24:16,160 --> 00:24:20,880 like integrated further, I mean either you're like doing the 384 00:24:20,880 --> 00:24:25,080 trading yourself or you're providing some service to 385 00:24:25,080 --> 00:24:27,720 someone that wants to sell transactions on chain, right. 386 00:24:28,880 --> 00:24:32,760 Let's say that you're a trader, right you're you're considering 387 00:24:33,200 --> 00:24:36,120 like running a block builder. It's not easy to run a block 388 00:24:36,120 --> 00:24:38,920 builder like not everyone can do it, not anyone can get access 389 00:24:38,920 --> 00:24:40,760 all the private order flow that's required. 390 00:24:41,640 --> 00:24:44,200 So you have like some, some cost, right, like you have some 391 00:24:44,440 --> 00:24:48,320 some potential revenue versus cost calculation that says like 392 00:24:48,320 --> 00:24:51,520 are we going to do trading or not or do block building or not. 393 00:24:52,600 --> 00:24:55,120 So they're they're willing to pay for someone else to provide 394 00:24:55,120 --> 00:24:57,960 that service if they don't have to sort of spin up themselves. 395 00:24:58,520 --> 00:25:03,160 I think this exists with several other systems as well where like 396 00:25:03,160 --> 00:25:07,920 they would be willing to to submit flow to the block builder 397 00:25:07,920 --> 00:25:10,800 if they can get some guarantees or like partnership on on how it 398 00:25:10,800 --> 00:25:13,440 works. I genuinely don't think block 399 00:25:13,440 --> 00:25:16,840 building alone is a very good business to run. 400 00:25:17,280 --> 00:25:21,760 It's like very competitive marketplace and unless you're 401 00:25:21,760 --> 00:25:28,990 able to get some some very like special deals, basically the 402 00:25:28,990 --> 00:25:33,790 profits get competed away. So the model is not necessarily 403 00:25:33,790 --> 00:25:38,670 super attractive. So what is going to be the kind 404 00:25:38,670 --> 00:25:41,670 of future of block building? So you think there's other 405 00:25:41,670 --> 00:25:46,870 products that are gonna be built by block builders and what do 406 00:25:46,870 --> 00:25:48,310 you think are the most promising ones? 407 00:25:48,310 --> 00:25:53,600 I think there's two trends that are worth keeping an eye on. 408 00:25:53,600 --> 00:26:00,000 One is a transition of order flow from from public venues to 409 00:26:00,080 --> 00:26:05,720 private to venues in particular, like just the transition of you 410 00:26:05,720 --> 00:26:09,760 to swap order flow from the main pool to you to swap X. 411 00:26:11,120 --> 00:26:14,360 That will influence a lot of different things. 412 00:26:14,400 --> 00:26:19,240 I think that will mean that searcher builders become less 413 00:26:19,440 --> 00:26:24,200 competitive and less important and neutral builders will become 414 00:26:24,200 --> 00:26:29,160 more more prominent. Hey can you explain why and and 415 00:26:29,160 --> 00:26:33,840 maybe just also explain what is the unit swap X cuz I think many 416 00:26:33,840 --> 00:26:39,000 people will not be. Can I use that Unisop X is like 417 00:26:39,000 --> 00:26:42,280 the rfq platform? Basically that's Unisop launched 418 00:26:43,520 --> 00:26:49,920 works similarly to the one inch fusion, similarly to Cal swap in 419 00:26:49,920 --> 00:26:53,780 some ways as well. It aims to use all sources of 420 00:26:53,780 --> 00:26:57,140 liquidity to be able to provide price improvements through 421 00:26:57,460 --> 00:27:02,300 running an auction with a bunch of professional private market 422 00:27:02,300 --> 00:27:06,220 makers for the right to settle a swap. 423 00:27:07,180 --> 00:27:12,740 So what this does is it transitions a lot of activity 424 00:27:12,740 --> 00:27:14,380 away from the public transaction pool. 425 00:27:15,220 --> 00:27:18,980 So we can say sort of goodbye to all the Santo Chang, which is 426 00:27:18,980 --> 00:27:24,700 great, great for users but also reduces sort of the opportunity 427 00:27:24,700 --> 00:27:27,580 space for those trades to take place. 428 00:27:28,060 --> 00:27:33,180 It makes providing liquidity as a passive LP and AMM's a lot 429 00:27:33,180 --> 00:27:37,700 more difficult because you have active market makers that have 430 00:27:37,700 --> 00:27:42,060 first look on all the retail orders and the chance to settle 431 00:27:42,060 --> 00:27:45,300 those. And so as a liquidity provider 432 00:27:45,300 --> 00:27:49,510 in AMM, you sort of just get all of the leftovers that for 433 00:27:49,510 --> 00:27:53,310 whatever reason the the professional market makers 434 00:27:53,310 --> 00:27:55,310 aren't settling. So you have to ask yourself like 435 00:27:55,310 --> 00:27:58,190 why aren't the professional market makers settling this and 436 00:27:58,190 --> 00:28:03,670 why do I, why am I as a retail, you know, do I believe that I'm 437 00:28:03,670 --> 00:28:05,790 making money off of off of this trade? 438 00:28:07,350 --> 00:28:09,190 The answer is like probably you're not. 439 00:28:09,910 --> 00:28:13,550 And so I expect there to be just like a decrease in the amount of 440 00:28:13,550 --> 00:28:20,350 liquidity that's available on, on, on layer one AMM's and more 441 00:28:20,350 --> 00:28:25,830 and more of that that trading to sort of settle on on this either 442 00:28:27,310 --> 00:28:31,710 private liquidity or layer two. This in turn like sort of 443 00:28:31,710 --> 00:28:34,910 reduces the arbitrage opportunity between the layer 444 00:28:34,910 --> 00:28:38,190 one and like other trading venues and if you have less 445 00:28:38,190 --> 00:28:42,590 liquidity that's available to be targeted, you will expose less 446 00:28:42,590 --> 00:28:45,250 value. Separately, there's like Unsoft 447 00:28:45,250 --> 00:28:48,690 before and like hooks and like other like mitigations are being 448 00:28:48,810 --> 00:28:52,450 developed to be able to reduce the amount of LVR that are 449 00:28:52,450 --> 00:28:55,610 exposed by passive Lps. Essentially all these 450 00:28:55,610 --> 00:28:59,650 microstructure changes that you swap is introducing reduce the 451 00:28:59,650 --> 00:29:02,650 amount of MV that gets exposed in the system and reduce the 452 00:29:02,770 --> 00:29:07,250 opportunity for integrated searcher builders to sort of you 453 00:29:07,250 --> 00:29:10,290 know have a significant portion of the value of the blocks. 454 00:29:11,250 --> 00:29:15,090 So more and more of I think the value of the blocks is going to 455 00:29:15,090 --> 00:29:21,850 be distributed towards other use cases relative to just 456 00:29:22,050 --> 00:29:25,570 arbitraging you swap that will cause the shift away from 457 00:29:26,610 --> 00:29:28,090 searching builders to respond neutral builder. 458 00:29:29,810 --> 00:29:33,410 And then how do like neutral builders differentiate? 459 00:29:33,410 --> 00:29:35,930 I guess maybe we went into a bit, but is it mostly on this 460 00:29:35,930 --> 00:29:40,490 like SO offers they can make for? 461 00:29:41,000 --> 00:29:43,520 Private order flow? Or what other dimensions are 462 00:29:43,520 --> 00:29:45,640 there to differentiate for neutral building? 463 00:29:46,680 --> 00:29:47,880 I don't really think that there is. 464 00:29:47,880 --> 00:29:52,240 I think it's like a knife fight. Just do it better. 465 00:29:54,120 --> 00:30:01,640 And what are your thoughts on decentralizing? 466 00:30:02,300 --> 00:30:04,820 The block builder is this. Do you think this is? 467 00:30:05,580 --> 00:30:08,420 Is it possible to build a decentralized block builder that 468 00:30:08,420 --> 00:30:11,500 would be competitive and is there a reason to do it? 469 00:30:14,580 --> 00:30:19,020 I'd say it's. I mean I don't know that much I 470 00:30:19,020 --> 00:30:21,420 will say about what decentralized block building 471 00:30:21,540 --> 00:30:26,180 actually looks like and the designs that I've seen or the 472 00:30:26,180 --> 00:30:28,900 attempts at it have been unconvincing. 473 00:30:29,840 --> 00:30:33,160 There's like this bigger question mark which is like what 474 00:30:33,160 --> 00:30:35,960 does decentralization actually mean that like needs to be 475 00:30:35,960 --> 00:30:39,080 answered when you're even like trying to tackle these questions 476 00:30:39,480 --> 00:30:43,600 and the answers I've seen haven't necessarily made much 477 00:30:43,600 --> 00:30:46,400 sense to me. This is like a very deep rabbit 478 00:30:46,400 --> 00:30:51,400 hole and to like dive dive into. I I see all of these things as 479 00:30:51,400 --> 00:30:55,360 like games that you that you're developing and and the main 480 00:30:55,360 --> 00:31:00,840 metric to me about how decentralized it is is what are 481 00:31:00,840 --> 00:31:08,600 the barriers to entry and the advantages to like to scale or 482 00:31:08,600 --> 00:31:13,120 to network effects that the system involves. 483 00:31:13,720 --> 00:31:18,800 And I think the definitions and like proposals for decentralized 484 00:31:18,800 --> 00:31:22,160 block buildings to me are like trading off what is like a 485 00:31:22,160 --> 00:31:26,280 competitive market with individual agents for like a 486 00:31:26,280 --> 00:31:27,920 network that has network effects. 487 00:31:27,960 --> 00:31:30,040 And to me that doesn't necessarily feel more 488 00:31:30,040 --> 00:31:34,440 decentralized. Maybe there's there's something 489 00:31:34,440 --> 00:31:37,960 I'm missing, but it is susceptible to same 490 00:31:38,280 --> 00:31:41,040 centralization pressures at different nodes of the network 491 00:31:41,040 --> 00:31:45,960 as any participant in the latency system have. 492 00:31:47,160 --> 00:31:52,960 And so while it may have properties and look like a 493 00:31:53,400 --> 00:31:57,440 network, it doesn't necessarily seem very decentralized to me. 494 00:31:57,940 --> 00:31:59,940 So I don't know, I don't know where decentralized block 495 00:31:59,940 --> 00:32:02,660 building is is headed. I think there'll be a lot of 496 00:32:02,660 --> 00:32:05,820 experiments and attempts at developing it. 497 00:32:07,100 --> 00:32:11,300 I would tend to believe that there's there's something 498 00:32:11,300 --> 00:32:15,740 promising I think to be said for adding more constraints to the 499 00:32:15,740 --> 00:32:20,180 way that block building works and essentially giving 500 00:32:20,180 --> 00:32:29,580 validators the more control over the policies of what they accept 501 00:32:29,700 --> 00:32:32,340 as blocks. And maybe you could say that 502 00:32:32,340 --> 00:32:35,740 sort of decentralizes the the role of block building a little 503 00:32:35,740 --> 00:32:40,980 bit by constraining them to some set of rules that are enforced 504 00:32:40,980 --> 00:32:46,140 at the validator level. And so you are leveraging sort 505 00:32:46,140 --> 00:32:49,940 of more distributed validator set and the request that those 506 00:32:49,940 --> 00:32:54,830 validators may have while still having the block builder role to 507 00:32:55,470 --> 00:33:01,030 actually enact those requests. That might be like a direction 508 00:33:01,030 --> 00:33:05,430 for more trust list block building. 509 00:33:05,990 --> 00:33:08,550 But yeah, is there like a specific approach to 510 00:33:08,550 --> 00:33:10,150 decentralized block building that you've looked at? 511 00:33:10,830 --> 00:33:16,510 I mean it's I guess there's a few reasons why this. 512 00:33:16,670 --> 00:33:19,350 I have not looked at this in any depth but. 513 00:33:20,910 --> 00:33:23,870 But there is a some, some things that sort of like triggered this 514 00:33:23,870 --> 00:33:26,550 a little bit. I mean one thing is that you 515 00:33:26,550 --> 00:33:29,710 know there was like so much effort and thinking that kind of 516 00:33:29,710 --> 00:33:33,190 went into all let's try to make Ethereum, you know as 517 00:33:33,190 --> 00:33:35,910 decentralized as we can that there's like lots of different 518 00:33:35,910 --> 00:33:39,310 validators now did that work or not versus like other proofs 519 00:33:39,310 --> 00:33:41,430 they change. You can debate, right? 520 00:33:41,430 --> 00:33:43,430 There's some aspects in which theorem is maybe more 521 00:33:43,430 --> 00:33:47,310 decentralized, some aspects maybe it's like similar to other 522 00:33:47,310 --> 00:33:49,650 networks, but you know. It's still, it's pretty 523 00:33:49,650 --> 00:33:52,490 centralized, right. When you look at the that would 524 00:33:52,490 --> 00:33:55,810 say the validation level and then, but now we've had the 525 00:33:55,810 --> 00:33:58,370 segregation of taking the roles apart and then you have this 526 00:33:58,370 --> 00:34:00,170 block building. And all of a sudden if you look 527 00:34:00,170 --> 00:34:06,050 at this block building, well, it's sort of like you have a few 528 00:34:06,050 --> 00:34:11,690 block builders that produce most of the blocks and as you said, 529 00:34:11,690 --> 00:34:16,010 right I guess at the moment. The ones that are kind of 530 00:34:16,010 --> 00:34:19,290 dominant there is the search of builders, which means they're 531 00:34:20,090 --> 00:34:25,010 also looking for what is the optimal transaction ordering and 532 00:34:25,010 --> 00:34:28,250 finding their own things. And the ones best able to do 533 00:34:28,250 --> 00:34:29,969 that are kind of trading companies, right? 534 00:34:29,969 --> 00:34:31,650 They're also trade intenterized exchanges. 535 00:34:31,650 --> 00:34:36,330 So from just the sort of okay, what do we want this system to 536 00:34:36,330 --> 00:34:40,480 look like? Having a few companies like that 537 00:34:40,480 --> 00:34:43,639 that in the end have this critical role maybe doesn't seem 538 00:34:43,639 --> 00:34:47,120 like the most desirable outcome. So this is like one thing and 539 00:34:47,120 --> 00:34:52,679 then the other thing has to do with you know some people 540 00:34:52,679 --> 00:34:55,239 talking about. And I think this is also true 541 00:34:55,239 --> 00:34:58,680 for faith builder if my if my understanding is correct or 542 00:34:58,680 --> 00:35:02,160 being like a neutral builder right where? 543 00:35:02,590 --> 00:35:05,990 Were basically okay, let's say there's like some builders and 544 00:35:05,990 --> 00:35:09,270 they kind of maybe use the information against you when you 545 00:35:09,270 --> 00:35:11,110 send it to them. But then there are others and 546 00:35:11,110 --> 00:35:13,310 they say we promise we don't do that, right? 547 00:35:13,310 --> 00:35:17,110 Like so trust us. Of course, that also seems like 548 00:35:17,110 --> 00:35:19,670 a bizarre thing, right? Because all of a sudden the 549 00:35:19,670 --> 00:35:22,910 whole thing is about building like trustless systems, right, 550 00:35:22,910 --> 00:35:26,390 with crypto and and cryptography and you know decent trans 551 00:35:26,390 --> 00:35:30,230 network and stuff. So if then you have at the core 552 00:35:30,230 --> 00:35:34,200 of this whole thing. You have these block builders 553 00:35:34,200 --> 00:35:39,720 who, you know, claim that they do things with a certain, you 554 00:35:39,720 --> 00:35:42,280 know, following a certain process, and that's why people 555 00:35:42,280 --> 00:35:45,000 use them. But actually it's based on 556 00:35:45,000 --> 00:35:48,600 trust, and you can't, like, verify that you don't have, 557 00:35:48,600 --> 00:35:51,000 like, the kind of guarantees that we generally have in crypto 558 00:35:51,000 --> 00:35:52,880 networks. That also seems a little bit 559 00:35:52,880 --> 00:35:56,160 weird, and it seems like the kind of thing that often you be 560 00:35:56,160 --> 00:36:00,000 like, oh. Let us replace this trusted 561 00:36:00,000 --> 00:36:02,800 thing of the neutral block builder with something that's, 562 00:36:02,800 --> 00:36:06,440 you know, actually trustless. And then of course, 563 00:36:06,440 --> 00:36:11,880 decentralizing it is often is, I guess, generally the way that 564 00:36:11,880 --> 00:36:16,200 this is approached. I do think that there's a 565 00:36:16,200 --> 00:36:18,480 difference there between providing better trust 566 00:36:18,480 --> 00:36:21,640 guarantees and decentralizing. So like you could do block 567 00:36:21,640 --> 00:36:26,320 building in a centralized way while still providing strong 568 00:36:26,320 --> 00:36:31,570 privacy guarantees, right? And so you can have this what is 569 00:36:31,570 --> 00:36:34,090 like filled through trust with like hey, I won't front run you 570 00:36:34,090 --> 00:36:37,570 I promise be enforced cryptographically while still 571 00:36:37,570 --> 00:36:40,770 being operated in a centralized manner. 572 00:36:42,050 --> 00:36:45,050 I think that's the natural evolution that things will will 573 00:36:45,050 --> 00:36:48,090 go towards. I think the like decentralized 574 00:36:48,090 --> 00:36:52,490 component of it is the part that I'm I'm not seeing how it'll how 575 00:36:52,490 --> 00:36:55,810 it'll take place. The other like question is like, 576 00:36:56,580 --> 00:37:00,660 so let's say that we decide that we don't like that there's five 577 00:37:00,660 --> 00:37:05,060 whatever block builders and we don't like that there's sort of 578 00:37:05,060 --> 00:37:08,340 a you know a power law distribution and dominance 579 00:37:08,340 --> 00:37:10,940 between them. We want block building to be 580 00:37:10,940 --> 00:37:14,580 uniform distribution. The way to do that is by 581 00:37:14,580 --> 00:37:16,460 creating centralization somewhere else. 582 00:37:16,780 --> 00:37:20,300 You say, OK, then you put all of the order flow within a single 583 00:37:20,300 --> 00:37:24,250 system and you sort of selectively distribute it to 584 00:37:24,250 --> 00:37:27,050 various from block builders and then force the fact that this 585 00:37:27,050 --> 00:37:29,490 uniform distribution happens on the block building side. 586 00:37:30,130 --> 00:37:33,170 You've just introduced the problem at a different layer, 587 00:37:33,170 --> 00:37:35,570 which is the order flow aggregation layer. 588 00:37:36,050 --> 00:37:40,170 And then are you happy to have as like you know, 2-3 different 589 00:37:40,170 --> 00:37:43,810 order flow systems that process all of the order flow And 590 00:37:44,330 --> 00:37:48,850 Ethereum, is that better? Is that like a meaningful 591 00:37:48,850 --> 00:37:52,660 improvement? Yeah, I think those those like 592 00:37:52,660 --> 00:37:56,820 end up being the directions of inquiry that that I get, that I 593 00:37:56,820 --> 00:38:01,300 get concerned about. I we also wanted to talk I guess 594 00:38:01,300 --> 00:38:06,700 about the PBS system in general. Is this the right direction and 595 00:38:06,700 --> 00:38:08,100 your thoughts? I guess we already covered it a 596 00:38:08,100 --> 00:38:12,140 bit, but given you know, Bright and me both like say there's 597 00:38:12,140 --> 00:38:14,020 like different chains and there's like different 598 00:38:14,020 --> 00:38:16,060 approaches to how MVV is treated, I think. 599 00:38:17,010 --> 00:38:20,050 Obviously if you're almost sort of the 1st and still like the 600 00:38:20,050 --> 00:38:24,770 main venue where the liquidity is and everything and yeah, we 601 00:38:24,770 --> 00:38:28,410 wanted to capture your thoughts on like would you do it again 602 00:38:28,410 --> 00:38:32,250 like this or is PBS the right way or should it be done 603 00:38:32,250 --> 00:38:36,130 differently on other chains? So for example, you know in 604 00:38:36,130 --> 00:38:40,730 Cosmos there is like different ways to try and handle MVV. 605 00:38:41,610 --> 00:38:44,170 I don't know how familiar you are with them, but like, yeah, I 606 00:38:44,170 --> 00:38:45,170 guess. Just wanted to hear your 607 00:38:45,170 --> 00:38:47,700 thoughts on. Is PBS the right direction for 608 00:38:47,700 --> 00:38:50,980 for proof stake that works or are there other ways to to 609 00:38:50,980 --> 00:38:52,900 handle this? I don't know. 610 00:38:54,900 --> 00:38:56,700 I don't know. I think it was right to 611 00:38:56,740 --> 00:38:58,620 Etherium. I think it was right for 612 00:38:58,620 --> 00:39:01,420 Ethereum at the merge like they needed enough boost and like 613 00:39:01,420 --> 00:39:06,220 that was that was probably good. Like they need an MV solution at 614 00:39:06,220 --> 00:39:09,020 that time and it worked out. Is it right for Ethereum in the 615 00:39:09,020 --> 00:39:10,900 future? Is it right for other chains? 616 00:39:10,900 --> 00:39:14,540 Is it right for layer twos? It's not clear to me. 617 00:39:15,690 --> 00:39:17,450 I mean, you guys have looked at this a lot, right? 618 00:39:17,570 --> 00:39:22,810 What do you guys think? So if you look at like Cosmos 619 00:39:22,810 --> 00:39:32,170 chains, then it seems like there are way easier ways to address 620 00:39:32,170 --> 00:39:34,410 this issue. I mean, I think in the Cosmos 621 00:39:34,410 --> 00:39:36,850 chain, right? You already have the situation 622 00:39:36,850 --> 00:39:42,730 that you have validators. And the validators depend on the 623 00:39:42,730 --> 00:39:46,650 token holders to stay with them. And the token holders of course 624 00:39:46,650 --> 00:39:50,410 want this chain to work properly, they want the token to 625 00:39:50,410 --> 00:39:54,210 do well. So if now you have some 626 00:39:54,210 --> 00:40:00,050 validator that says I'm going to sandwich, let's say you have 627 00:40:00,050 --> 00:40:03,250 some sort of decks chain, and then some validator says I'm 628 00:40:03,250 --> 00:40:06,700 going to sandwich everyone. Then the then the token, all 629 00:40:06,700 --> 00:40:09,300 those are like, what the hell. This is like undermining my 630 00:40:09,300 --> 00:40:11,300 project. I'm not going to stick with you. 631 00:40:11,300 --> 00:40:13,100 Right. So that this is like one thing. 632 00:40:14,380 --> 00:40:18,140 So I think that makes it like way harder for someone to say, 633 00:40:18,140 --> 00:40:21,060 oh, I'm going to run this kind of strategy that you have on 634 00:40:21,060 --> 00:40:23,180 etheorem, right? Because any theorem, I think you 635 00:40:23,180 --> 00:40:24,660 just have a very different landscape. 636 00:40:25,220 --> 00:40:27,900 So I think that's one thing. And yeah, I guess you've had a 637 00:40:27,900 --> 00:40:30,500 few different experiments on on Cosmos, right? 638 00:40:30,500 --> 00:40:34,500 Like like for examples Moses where they have basically some 639 00:40:34,500 --> 00:40:37,780 in protocol MEV capture where they're you know they're 640 00:40:37,780 --> 00:40:42,940 basically running some Mev bots and then it goes kind of goes 641 00:40:42,980 --> 00:40:48,180 back to the stakers and mostly and then and then I guess I 642 00:40:48,180 --> 00:40:53,820 think in dydx right where we have, we've done this report 643 00:40:54,620 --> 00:40:58,700 where there's a committee that or where it's like okay, you 644 00:40:58,700 --> 00:41:03,760 cannot the validators are not allowed right to do basically 645 00:41:03,760 --> 00:41:09,440 MEV and then there would be some kind of taking looking at data 646 00:41:09,560 --> 00:41:12,560 and trying to detect things and then the ability to have 647 00:41:12,800 --> 00:41:15,840 slashing of validators who do it anyway. 648 00:41:16,400 --> 00:41:19,160 So these things feel like pretty straightforward and easy 649 00:41:19,160 --> 00:41:23,920 solution that actually solved the problem pretty well for the 650 00:41:23,920 --> 00:41:25,880 foreseeable future. Maybe at some point it will 651 00:41:25,880 --> 00:41:29,320 break down in some ways, but there it seems like proposal 652 00:41:29,320 --> 00:41:32,400 build a separation, it adds a huge amount of complexity. 653 00:41:33,560 --> 00:41:37,160 It doesn't really seem to. Does it solve the core problems 654 00:41:37,160 --> 00:41:39,240 any better than that? I don't think so. 655 00:41:39,800 --> 00:41:43,760 So doesn't feel like the right way for at least for let's say 656 00:41:43,760 --> 00:41:46,840 Cosmos change now. OK, you have other things again 657 00:41:47,120 --> 00:41:51,320 Solana, I don't know difficult. It's again very different. 658 00:41:51,320 --> 00:41:55,670 So yeah. I think the one thing that's 659 00:41:55,670 --> 00:42:01,830 that PBS is, is sort of validator is rejecting their 660 00:42:01,830 --> 00:42:04,550 power, right. In many ways like they're saying 661 00:42:04,550 --> 00:42:09,390 like all these systems of that are that do like leader 662 00:42:09,390 --> 00:42:11,470 selection. They essentially say, well, 663 00:42:11,470 --> 00:42:14,030 there you have, you have power, you have all the power and like 664 00:42:14,030 --> 00:42:18,070 you have to like figure out what to do with it And and PBS is 665 00:42:18,070 --> 00:42:20,550 kind of like rejection of this power, Say like, oh like no. 666 00:42:21,450 --> 00:42:24,250 We don't want to use this power, but we want to get paid for it 667 00:42:24,290 --> 00:42:27,210 and we're just going to like give this power to the highest 668 00:42:27,210 --> 00:42:30,050 bidder, right? That's sort of that's sort of 669 00:42:30,210 --> 00:42:32,370 what it says. But is that actually what's like 670 00:42:32,450 --> 00:42:35,410 Max, you know, utility for the end user? 671 00:42:35,530 --> 00:42:39,290 Or is there some other approach to dealing with the power that 672 00:42:39,290 --> 00:42:42,250 these these validators receive and and the leader selection 673 00:42:42,250 --> 00:42:47,570 process that provides guarantees are more useful for friend users 674 00:42:47,930 --> 00:42:50,080 and that's. You know, that's how I would 675 00:42:50,080 --> 00:42:55,400 frame what you're discussing. If validators opt for 676 00:42:55,400 --> 00:42:58,680 constraining the types of activity or types of blocks that 677 00:42:58,680 --> 00:43:03,840 they produce based off of some, you know, slashing mechanism or 678 00:43:03,840 --> 00:43:09,000 whatever else, you may be able to generate more more utility 679 00:43:09,000 --> 00:43:13,160 for the end users, whether it's first in, first out system or 680 00:43:15,320 --> 00:43:18,770 or. Other sort of rules, I mean even 681 00:43:18,770 --> 00:43:21,570 censorship rules at the end of the day, are policies that are 682 00:43:21,570 --> 00:43:26,130 enforced by the validators. And so you can come up with sort 683 00:43:26,130 --> 00:43:29,570 of a framework for thinking of policies that the validators 684 00:43:29,570 --> 00:43:34,770 subscribe to and offer to the consumers of the block space 685 00:43:34,770 --> 00:43:37,610 that they offer. So that was also what you 686 00:43:37,610 --> 00:43:40,650 mentioned earlier that this way if you would have this sort of 687 00:43:40,650 --> 00:43:42,900 system. Like different block builders 688 00:43:42,900 --> 00:43:45,740 could serve different constraints for the validators 689 00:43:45,740 --> 00:43:48,100 and that way essentialize more. Is that correct? 690 00:43:48,660 --> 00:43:51,820 Yeah, exactly. Yeah, are interesting. 691 00:43:52,260 --> 00:43:55,620 I do think that there was, there's this like tendency, 692 00:43:55,620 --> 00:43:59,500 right, like Brian what you said with the text chain and people 693 00:43:59,500 --> 00:44:02,820 saying if you said which everyone you're destroying the 694 00:44:02,820 --> 00:44:04,500 project. But I guess at the same time if 695 00:44:04,500 --> 00:44:06,980 you delegate to that validator who does it, you might get a 696 00:44:06,980 --> 00:44:09,060 higher yield, right? So I guess there's like a sort 697 00:44:09,060 --> 00:44:12,050 of. Issue there where if you just do 698 00:44:12,050 --> 00:44:15,450 it and delegate to this validator it probably you just 699 00:44:15,450 --> 00:44:18,050 benefit and the chain doesn't like go to 0. 700 00:44:18,050 --> 00:44:23,450 But yeah, I was wondering if that maybe is what drives it. 701 00:44:23,850 --> 00:44:27,810 I mean you could have, you could of course have, yeah, say hey, 702 00:44:27,810 --> 00:44:33,250 I'm gonna launch a validator and I'm gonna be like maximization 703 00:44:33,250 --> 00:44:37,050 validator and maybe you can generate higher returns and 704 00:44:37,050 --> 00:44:40,740 maybe people stay with you. It's just the challenge is if 705 00:44:40,740 --> 00:44:45,860 you have, if you have like governance, I mean it is very 706 00:44:45,860 --> 00:44:48,460 easy then for someone to say, hey, we're going to make a 707 00:44:48,740 --> 00:44:52,460 governance proposal and we're going to say like hey this this 708 00:44:52,460 --> 00:44:55,420 guy here, he's like messing up our project. 709 00:44:55,420 --> 00:44:58,940 So I don't know, we just slash them or we say that's not 710 00:44:58,940 --> 00:45:05,300 allowed or we we take some action that goes against that. 711 00:45:06,300 --> 00:45:08,420 I mean, I think this is maybe a little bit. 712 00:45:09,560 --> 00:45:15,240 I guess that's also where the whole app chain world can be a 713 00:45:15,240 --> 00:45:18,240 bit different from the etherum world, right? 714 00:45:18,240 --> 00:45:21,320 Where the etherum world is really doesn't really allow for 715 00:45:21,320 --> 00:45:25,240 that and pros and cons on that, right? 716 00:45:25,240 --> 00:45:27,960 This, of course, it tells us its advantages. 717 00:45:28,400 --> 00:45:33,920 Have you looked at Mev burn and do you have thoughts on it and 718 00:45:34,400 --> 00:45:36,560 we were gonna ask you for your faults on it. 719 00:45:39,160 --> 00:45:46,680 I don't really have I mean I I am like we're at the at the 720 00:45:46,680 --> 00:45:52,280 highest level of sort of like what it its intention is but I I 721 00:45:52,280 --> 00:45:55,360 don't have I yeah it would be great if you could explain first 722 00:45:55,360 --> 00:45:58,600 of all what MV burn is and then what your faults are. 723 00:45:59,510 --> 00:46:03,830 So one of the benefits of PBS is that you have a public price for 724 00:46:03,830 --> 00:46:07,350 the value of blocks that are being produced, right, which is 725 00:46:07,350 --> 00:46:10,830 not something that existed in a pre MEF, boost, pre PBS world. 726 00:46:12,230 --> 00:46:16,310 Prior to that you sort, you could see the amount of 727 00:46:16,310 --> 00:46:19,830 transfers to Coinbase that are taking place and you could like 728 00:46:19,910 --> 00:46:23,870 you know sum it I guess and like see out of bounds like what what 729 00:46:23,870 --> 00:46:26,230 it looks like. But that wasn't the price the 730 00:46:26,230 --> 00:46:28,430 venue for for the price discovery to take place. 731 00:46:28,790 --> 00:46:31,370 Right like. If a minor wanted to have a 732 00:46:31,370 --> 00:46:34,570 different approach to collecting fees, like they could do so and 733 00:46:34,570 --> 00:46:36,970 it wouldn't reflect the value of the blocks that they produce. 734 00:46:37,690 --> 00:46:41,490 Similarly, we sort of saw a lot of miners say like we're only 735 00:46:41,490 --> 00:46:47,250 disputing like block rewards to to the miners in our mining pool 736 00:46:47,330 --> 00:46:51,650 and we're keeping all of the MEV rewards to ourselves as like our 737 00:46:51,850 --> 00:46:54,170 monetization. And in some ways like that was 738 00:46:54,170 --> 00:46:56,410 possible because the miners didn't see sort of this one 739 00:46:56,410 --> 00:46:58,130 number that said like, OK, here's like. 740 00:46:58,650 --> 00:47:02,650 The value of the block that I mined, you know, you just know 741 00:47:02,650 --> 00:47:05,570 that the blocker war that was included in that block. 742 00:47:06,850 --> 00:47:11,690 So now we have this public value for the blocks and the question 743 00:47:11,690 --> 00:47:14,730 is like how does this value get distributed? 744 00:47:15,290 --> 00:47:22,050 Right now it's sort of left to the mining pool operators or 745 00:47:22,130 --> 00:47:25,850 node infrastructure providers to decide how they wanna distribute 746 00:47:25,850 --> 00:47:29,860 that to their constituents. Most of the time it's they keep 747 00:47:29,860 --> 00:47:33,740 a percentage of those rewards and like distribute the rest. 748 00:47:34,300 --> 00:47:39,100 There's still some incentive to optimize this in some way and 749 00:47:39,260 --> 00:47:41,260 whatever else. And there's more and more 750 00:47:41,260 --> 00:47:45,180 discussion about maybe this value should be redistributed to 751 00:47:45,180 --> 00:47:49,180 the network and not necessarily to validators. 752 00:47:49,540 --> 00:47:52,980 So Mev Byrne is essentially saying, hey, we see this public 753 00:47:52,980 --> 00:47:55,690 value. Is there a way to enforce that 754 00:47:55,690 --> 00:48:02,890 this gets burned similar to the AP1559 rather than distributed 755 00:48:03,090 --> 00:48:07,890 to to validators? There's like interest from token 756 00:48:07,890 --> 00:48:13,010 holders like ETH holders to do this, and I think there's some 757 00:48:13,010 --> 00:48:16,250 arguments for like network stabilities and stability as to 758 00:48:16,250 --> 00:48:21,810 why why why this should be done. From my view, it seems to be net 759 00:48:21,810 --> 00:48:24,320 negative for. For anyone who runs a validator 760 00:48:24,600 --> 00:48:26,960 in terms of the yield that they're going to be able to 761 00:48:26,960 --> 00:48:29,960 generate. Yeah, it's interesting. 762 00:48:29,960 --> 00:48:33,400 It seems it's beneficial for ETH holders that don't stake right 763 00:48:33,400 --> 00:48:37,680 and maybe discourages staking since the yield will be much 764 00:48:37,680 --> 00:48:40,960 lower. And I guess I personally also 765 00:48:40,960 --> 00:48:45,280 probably saw maybe as like one of the revenue sources for 766 00:48:45,840 --> 00:48:51,960 moving away from inflationary token increase, so. 767 00:48:52,660 --> 00:48:56,020 When I think it is similar to what you know I guess in osmosis 768 00:48:56,020 --> 00:48:59,740 or these models where essentially the map is captured 769 00:48:59,740 --> 00:49:02,180 by the chain and like redistributed to the token 770 00:49:02,180 --> 00:49:05,380 itself. So I think it's important that 771 00:49:05,380 --> 00:49:10,180 like you know people do stake and infrastructure providers do 772 00:49:10,180 --> 00:49:12,740 run notes. So like somewhere this revenue 773 00:49:12,740 --> 00:49:15,580 has to come from and probably maybe is like a pretty good 774 00:49:15,620 --> 00:49:20,060 place for that. Yeah, I guess the questions also 775 00:49:21,110 --> 00:49:23,710 and I think you'll have a much better understanding of this. 776 00:49:23,710 --> 00:49:28,110 But, you know, if something like MV burn was created, you know, 777 00:49:28,110 --> 00:49:32,430 are there ways to kind of route around that, you know, to 778 00:49:32,430 --> 00:49:36,870 through some sort of, you know, off chain, like basically trying 779 00:49:36,870 --> 00:49:40,430 to pretend to the chain that this block doesn't have a lot of 780 00:49:40,430 --> 00:49:43,470 values, that little is burned, but actually it has more value 781 00:49:43,470 --> 00:49:47,510 that than is distributed in some way between block builders, 782 00:49:47,510 --> 00:49:49,230 validators in some way. Yeah. 783 00:49:50,530 --> 00:49:55,690 What do you focus on that? Yes, I mean both EAP 5059 and 784 00:49:55,770 --> 00:50:02,650 Mev Byrne to me is increasing the potential value of collusion 785 00:50:03,290 --> 00:50:09,610 is like a way to think about it. These mechanisms all have some 786 00:50:09,610 --> 00:50:15,130 sort of assumption that there's some price to collusion that's 787 00:50:15,130 --> 00:50:17,090 higher than the benefit of it and. 788 00:50:17,580 --> 00:50:21,380 You know, therefore the collusion isn't going to to take 789 00:50:21,380 --> 00:50:23,380 place. It definitely seems to me that 790 00:50:23,380 --> 00:50:27,580 with Mel Byrne there's a pretty strong incentive to attempt to 791 00:50:28,300 --> 00:50:31,500 bypass it. Again, Mel Byrne is like there's 792 00:50:31,500 --> 00:50:36,500 like a few ideas for how it's it's implemented and it depends 793 00:50:36,500 --> 00:50:39,780 on the specifics of it and the model that does deployed. 794 00:50:41,460 --> 00:50:45,380 But the incentive is there is is it significant enough to cause a 795 00:50:45,380 --> 00:50:50,230 change in behavior? Given both economic incentive 796 00:50:50,350 --> 00:50:54,990 and social political incentive. I don't. 797 00:50:55,870 --> 00:50:57,870 Know to me it's still like what does it solve? 798 00:50:57,950 --> 00:50:59,430 What does Matt Bern actually solve? 799 00:50:59,750 --> 00:51:03,870 It's not clear. I mean, I guess one thing I've 800 00:51:03,870 --> 00:51:12,200 heard was more that, well if Mev that it smooths, you know that 801 00:51:12,240 --> 00:51:16,200 you don't have like huge returns being made on some blocks and 802 00:51:16,200 --> 00:51:20,160 that then this sort of incentivizes more concentration 803 00:51:20,200 --> 00:51:23,640 and staking pools or something. Because like if you just run a 804 00:51:23,680 --> 00:51:28,080 few validators, then you know it's going to be very rare that 805 00:51:28,080 --> 00:51:31,480 you earn like a lot of money from MVV and and so. 806 00:51:32,520 --> 00:51:36,560 But I don't know, that's not sure that's how convincing the 807 00:51:36,560 --> 00:51:39,960 argument is to me. This moves it by. 808 00:51:40,850 --> 00:51:46,810 Removing it right. So it's like it's like an anti 809 00:51:46,810 --> 00:51:49,970 staking pool thing. I heard that as one. 810 00:51:49,970 --> 00:51:54,530 I think maybe Justin Drake gave that as one explanation for it. 811 00:51:55,370 --> 00:52:00,890 But yeah, I think if you sort of look at it on a high level or 812 00:52:00,890 --> 00:52:04,850 maybe we should talk more about that actually if you listen area 813 00:52:04,850 --> 00:52:07,810 we haven't really talked about, we'd love to dive into a bit. 814 00:52:08,290 --> 00:52:11,770 And so right now we talked about block builders and how they are 815 00:52:11,770 --> 00:52:13,730 basically capturing a bunch of value. 816 00:52:14,610 --> 00:52:19,370 But of course we have maybe other places where this can 817 00:52:19,370 --> 00:52:23,490 happen too, right. So one would be the wallet or 818 00:52:23,490 --> 00:52:28,610 another one might be let's say some Ethereum app that a little 819 00:52:28,610 --> 00:52:33,290 bit like what UNISOP or Cal software doing to building some 820 00:52:33,290 --> 00:52:35,970 kind of thing that also circumvents the MUV. 821 00:52:36,980 --> 00:52:39,220 And I'm curious, how do you, what do you feel actually is 822 00:52:39,220 --> 00:52:42,460 gonna be the role of maybe wallets or some of these other 823 00:52:42,460 --> 00:52:45,900 players in the Mev supply network? 824 00:52:46,980 --> 00:52:51,460 My claim has always been that MEV is best mitigated at the 825 00:52:51,460 --> 00:52:54,820 application layer. If you build better applications 826 00:52:55,140 --> 00:53:01,220 that deal with MEV, you're able to provide better user 827 00:53:01,700 --> 00:53:06,340 guarantees and superior user experience because of that. 828 00:53:07,200 --> 00:53:10,640 And to me, a wallet and an application are one of the same, 829 00:53:11,240 --> 00:53:14,760 The venue that the user is interfacing with to to get their 830 00:53:15,800 --> 00:53:20,440 their preferences expressed the role. 831 00:53:20,840 --> 00:53:23,000 There's this like interesting role that the wallets have 832 00:53:23,000 --> 00:53:29,240 played historically, which is to be both the assigner and an RPC 833 00:53:29,240 --> 00:53:33,160 provider right? With like a nice UI around it. 834 00:53:34,860 --> 00:53:42,460 But not really try to provide the experience of what the user 835 00:53:42,460 --> 00:53:47,900 actually wants to do, right? Let's say that's the meta mask 836 00:53:47,900 --> 00:53:49,340 thing. We are a browser extension. 837 00:53:49,340 --> 00:53:50,700 We work on all the different apps. 838 00:53:51,460 --> 00:53:54,940 We are sort of agnostic to the type of activity that you do. 839 00:53:55,260 --> 00:53:58,100 All we do is we provide the signing, we provide the RPC 840 00:53:58,100 --> 00:54:01,860 service. A lot of Mev now is about maybe. 841 00:54:02,940 --> 00:54:04,940 The RPC service has become more complicated. 842 00:54:05,020 --> 00:54:07,500 There's all these different private venues through which you 843 00:54:07,500 --> 00:54:10,060 can route, and not just to the public transaction pool. 844 00:54:10,540 --> 00:54:14,380 The type of RPC service that you use and the guarantees that you 845 00:54:14,380 --> 00:54:18,300 get from them is more closely tied to the type of activity 846 00:54:18,300 --> 00:54:23,060 that the user is doing. And so as an application 847 00:54:23,060 --> 00:54:25,820 developer who really cares about the user experience and like 848 00:54:25,820 --> 00:54:28,260 really understand what the user's trying to achieve, I 849 00:54:28,260 --> 00:54:31,540 would be in a better position to understand which RPC to use. 850 00:54:32,320 --> 00:54:35,360 So how do I negotiate this with the fact that right now that 851 00:54:35,360 --> 00:54:39,240 service is being handled by some external product? 852 00:54:39,640 --> 00:54:43,080 I want to bring that level of control over the user experience 853 00:54:43,080 --> 00:54:46,440 within my application. There's a bit of a tension 854 00:54:46,440 --> 00:54:49,960 between applications and and wallets in this way. 855 00:54:50,240 --> 00:54:52,880 And so I think that the natural result is we'll see more sort of 856 00:54:52,880 --> 00:54:55,520 separation between these things like more applications launched 857 00:54:55,520 --> 00:54:59,840 in wallets, more wallets have an application feel and sort of 858 00:54:59,840 --> 00:55:04,110 integrate the full sort of user. Life cycle within their own, 859 00:55:04,670 --> 00:55:07,430 their own system. The role that they play in MEV 860 00:55:07,430 --> 00:55:12,110 is is the same, right? Like they want to provide the 861 00:55:12,110 --> 00:55:14,910 best guarantees to their users to retain them, increase the 862 00:55:14,910 --> 00:55:16,270 level of activity that they have. 863 00:55:16,710 --> 00:55:19,630 That could mean protecting them from MV, It could mean giving 864 00:55:19,630 --> 00:55:23,430 them better prices, it could mean selling their order flow 865 00:55:23,630 --> 00:55:27,350 for monetization reasons. There's a bunch of different 866 00:55:27,350 --> 00:55:31,070 things that that applications can do. 867 00:55:32,230 --> 00:55:35,710 Cool. I want to ask another top like 868 00:55:35,710 --> 00:55:40,110 one maybe one last topic here. So this is another area that 869 00:55:40,110 --> 00:55:42,950 does gets a lot of attention at the moment, which is intense. 870 00:55:43,310 --> 00:55:47,230 And I mean what are intense. We've before spoke about this 871 00:55:47,230 --> 00:55:50,070 RFQ thing, which I think is a great example of an intent, 872 00:55:50,070 --> 00:55:53,430 right, where it's basically someone saying instead of 873 00:55:53,430 --> 00:55:57,390 saying, you know, transaction that's like, you know, OK, do 874 00:55:57,390 --> 00:56:00,810 exactly this on chain, I guess mostly with trade. 875 00:56:00,810 --> 00:56:06,010 So I'd swap X for Y. You say like, hey, I want to, I 876 00:56:06,010 --> 00:56:10,090 want this kind of outcome. You define some sort of criteria 877 00:56:10,090 --> 00:56:14,570 for outcomes and then and then you know there will be a job for 878 00:56:14,570 --> 00:56:17,490 somebody else to figure out what is the best way of satisfying 879 00:56:17,490 --> 00:56:20,330 that outcome. And you know, there's some 880 00:56:20,330 --> 00:56:23,050 blockchains that are very much focusing on this. 881 00:56:23,050 --> 00:56:25,650 I guess like a Noma is one example, but some others and 882 00:56:25,650 --> 00:56:29,850 then of course there's intense applications doing things like 883 00:56:29,850 --> 00:56:32,770 that on Ethereum. Do you think intents are gonna 884 00:56:32,770 --> 00:56:35,730 have a major impact on this MEV supply network? 885 00:56:37,050 --> 00:56:39,850 Yes and no. I think intents are confusing 886 00:56:39,850 --> 00:56:44,490 because the way that it's being framed is as opening up like a 887 00:56:44,570 --> 00:56:48,970 very broad new set of possibilities in a way that I'm 888 00:56:48,970 --> 00:56:52,410 not sure how broad it is. There's a quip about intents, 889 00:56:52,410 --> 00:56:55,750 which is. An intent is no longer an intent 890 00:56:55,790 --> 00:57:00,310 once it has product market fit. So an intent is called like an 891 00:57:00,310 --> 00:57:06,110 RFQ if it's like a swapping use case, right? 892 00:57:06,470 --> 00:57:11,070 Or an intent is called like a bridge transaction if it's about 893 00:57:11,070 --> 00:57:16,550 going to a different chain. So I think intents where there 894 00:57:16,550 --> 00:57:21,390 is a need is a like more advanced transaction types. 895 00:57:21,940 --> 00:57:25,940 That allow for performing types of activities that you aren't 896 00:57:25,940 --> 00:57:30,660 able to perform otherwise, that constrain the execution space 897 00:57:30,660 --> 00:57:36,060 based off of things that are either inefficient or have 898 00:57:36,300 --> 00:57:39,900 difficult trust boundaries to enforce directly on a layer one, 899 00:57:40,540 --> 00:57:45,260 and those seem mostly like standards to me. 900 00:57:46,170 --> 00:57:49,650 An intent that's able to have conditional execution on 901 00:57:49,650 --> 00:57:55,570 multiple different chains and able to bring constraints on 902 00:57:55,610 --> 00:58:02,090 external state is fairly interesting as sort of a model. 903 00:58:02,810 --> 00:58:06,930 Whether or it will be developed in a generalized case or a 904 00:58:06,930 --> 00:58:10,770 specific case to me is sort of a key question. 905 00:58:11,530 --> 00:58:14,700 So. In this sort of infinite games, 906 00:58:15,020 --> 00:58:18,700 transaction supply, network framing of things. 907 00:58:19,460 --> 00:58:21,420 All these different transaction types, all these different 908 00:58:21,420 --> 00:58:25,340 message types are being passed around are intense right? 909 00:58:25,700 --> 00:58:30,500 They are ways of communicating some preferences over what 910 00:58:30,500 --> 00:58:34,300 solvers on the other side are doing and how they're re 911 00:58:34,300 --> 00:58:39,540 aggregating those messages together, the exact format and 912 00:58:39,540 --> 00:58:42,090 like use cases. Are all of these different 913 00:58:42,090 --> 00:58:46,050 messages going to use the same standards for being expressed or 914 00:58:46,050 --> 00:58:48,610 are they going to be very use case specific? 915 00:58:49,450 --> 00:58:51,450 I would tend to believe that it's more likely to be use case 916 00:58:51,450 --> 00:58:54,690 specific. While like Intents can be like a 917 00:58:54,690 --> 00:58:57,530 very useful generalized framework for thinking about the 918 00:58:58,850 --> 00:59:01,970 expressivity of these messages and sort of pushing for more 919 00:59:01,970 --> 00:59:06,370 expressive systems, I think the development will be use case 920 00:59:06,370 --> 00:59:13,590 driven and it'll be about what. Is this more expressive message 921 00:59:13,590 --> 00:59:19,350 type enabling to build as an application rather than the 922 00:59:19,350 --> 00:59:20,950 message type being the core focus? 923 00:59:22,790 --> 00:59:25,230 Cool. Well, thank you so much Stefan 924 00:59:25,230 --> 00:59:28,270 for coming on. I think Mev feels like one of 925 00:59:28,270 --> 00:59:33,190 these rabbit holes that's just like complex and changing at a 926 00:59:33,190 --> 00:59:38,690 really rapid rate and also really fundamental for how all 927 00:59:38,690 --> 00:59:40,250 these systems are going to work in the future. 928 00:59:40,250 --> 00:59:43,330 So it's really great to, you know, have you back on and talk 929 00:59:43,330 --> 00:59:45,370 about these things and I'm excited about the work you're 930 00:59:45,370 --> 00:59:49,210 doing in Frontier and and all the all the things are going to 931 00:59:49,210 --> 00:59:51,170 come out of that. So thanks so much. 932 00:59:51,770 --> 00:59:53,810 Yeah, always a pleasure to come back on Hope. 933 00:59:53,810 --> 00:59:56,330 I had a few insights that were interesting. 934 00:59:57,010 --> 00:59:59,330 Absolutely cool. Thanks so much and thanks to the 935 00:59:59,330 --> 01:00:00,570 listeners. We look forward to being back 936 01:00:00,570 --> 01:00:04,410 next week. Thank you for joining us on this 937 01:00:04,410 --> 01:00:06,770 week's episode. We release new episodes every 938 01:00:06,770 --> 01:00:08,800 week. You can find and subscribe to 939 01:00:08,800 --> 01:00:12,560 the show on iTunes, Spotify, YouTube, SoundCloud, or wherever 940 01:00:12,560 --> 01:00:14,960 you listen to podcasts. And if you have a Google Home or 941 01:00:14,960 --> 01:00:17,680 Alexa device, you can tell it. To listen to the latest episode 942 01:00:17,680 --> 01:00:21,800 of the Epicenter podcast, go to epicenter.tv, subscribe for a 943 01:00:21,800 --> 01:00:23,440 full list of places where you can watch and listen. 944 01:00:23,440 --> 01:00:26,360 And while you're there, be sure to sign up for the newsletter so 945 01:00:26,360 --> 01:00:28,720 you get new episodes in your inbox as they're released. 946 01:00:28,720 --> 01:00:32,000 If you want to interact with us guests or other podcast 947 01:00:32,000 --> 01:00:33,640 listeners, you can follow us on Twitter. 948 01:00:34,040 --> 01:00:35,480 And please leave us a review on iTunes. 949 01:00:35,520 --> 01:00:38,080 It helps to be behind the show and we're always happy to read 950 01:00:38,080 --> 01:00:40,400 them. So thanks so much and we look 951 01:00:40,400 --> 01:00:41,520 forward to being back next week.