WEBVTT 00:09.177 --> 00:10.198 [SPEAKER_02]: What's up, everyone? 00:10.238 --> 00:12.721 [SPEAKER_02]: We're here, the crypto-mavage podcast. 00:12.861 --> 00:14.403 [SPEAKER_02]: And we have a great guest here for you today. 00:14.523 --> 00:15.604 [SPEAKER_02]: It's Ben Coverson. 00:15.644 --> 00:17.366 [SPEAKER_02]: He's the co-founder of Harmonic. 00:17.767 --> 00:24.554 [SPEAKER_02]: Harmonics and open block building infrastructure for Solana designed to make the network more fair, efficient, and competitive. 00:24.915 --> 00:26.897 [SPEAKER_02]: Ben really excited to have you on the podcast. 00:28.158 --> 00:28.939 [SPEAKER_00]: Yeah, excited to hear. 00:29.019 --> 00:30.240 [SPEAKER_00]: Thanks for having me. 00:30.260 --> 00:31.582 [SPEAKER_02]: Yeah, we're fired up to have you on. 00:31.602 --> 00:33.344 [SPEAKER_02]: We're going to talk a little bit about a harmonic. 00:33.364 --> 00:34.605 [SPEAKER_02]: We're going to talk Solana. 00:34.966 --> 00:37.969 [SPEAKER_02]: But before we do that, let's dive in a little bit about your background. 00:38.692 --> 00:39.593 [SPEAKER_00]: Yeah, yeah, sick. 00:39.614 --> 00:51.353 [SPEAKER_00]: So I started out in the HFT world, you know, while I was still in college in turn, at various firms, such as a CUNE capital and Citadel securities, you've probably heard of at least one of those. 00:52.955 --> 01:04.935 [SPEAKER_00]: They're all sorts of both like high performance computing work and PlongWork at firms became really, really interested in trading and like around the same time, you know, this is like early 21, late 2020. 01:05.017 --> 01:18.150 [SPEAKER_00]: There were, there was an increasing amount of people that were talking, especially the HFT world about like the Steve high stuff, probably like a pit of a time, was like reaching out to all the trading firms, like getting them to like provide data and stuff, right? 01:18.190 --> 01:24.676 [SPEAKER_00]: So like I invested in crypto, you know, and like I had continued to invest in crypto, you know, through like 2019, 2020, right? 01:24.896 --> 01:30.822 [SPEAKER_00]: Like I think had a small amount of salon at the time, but happened really played a 01:30.802 --> 01:35.537 [SPEAKER_00]: you know, with D-Fi as much until I started seeing this Oracle stuff in my college, it's like really interesting, right? 01:35.557 --> 01:42.840 [SPEAKER_00]: So like, I experimented with all the DAPS and, you know, I had like, on-chain order books like Serum, Rest in Peace, right? 01:43.041 --> 01:43.663 [SPEAKER_00]: Um, 01:44.099 --> 01:50.089 [SPEAKER_00]: I, when I've been transactions though, right, like when they actually landed the transactions were like extremely fast, right? 01:50.109 --> 01:59.525 [SPEAKER_00]: So like it was some time around then like late 2020 or early 2021 that I'm like, wow, like on chain finance, like DeFi, like HFT level DeFi, like might actually be possible. 01:59.545 --> 02:07.118 [SPEAKER_00]: Like there's, there's, there's a chance that these markets can can become, you know, as as competitive as centralized markets, right? 02:07.138 --> 02:10.123 [SPEAKER_00]: And that's, that's kind of the vision that I've always believed in, right, you know, temporal and 02:10.103 --> 02:29.142 [SPEAKER_00]: harmonic right and pretty much everything I've done since then has been like founded on this like principle they're like we we want to make defy more efficient than c5 right and only the last year we finally seen this started to see this come to fruition right so we have now on chain users are getting better fills on prop amms such as humidify 02:29.307 --> 02:32.433 [SPEAKER_00]: users are getting better fields than they're getting on Central Asia changes, right? 02:32.813 --> 02:35.598 [SPEAKER_00]: You have some Central Asia changes that take like 100 BIPs fees, right? 02:35.638 --> 02:42.711 [SPEAKER_00]: Others take 10 BIPs fees, all of these are much higher than the fills, that they're getting all these problems, right? 02:42.731 --> 02:46.718 [SPEAKER_00]: The problem is they're giving users better prices, better spreads, right? 02:47.439 --> 02:52.328 [SPEAKER_00]: It's pretty incredible to see this is like starting to come to fruition, right? 02:52.308 --> 03:09.388 [SPEAKER_00]: So like my vision right is for like Solana to be the chain to trade everything right and it is now the best chain to trade Solana itself right and is becoming the best chain to trade other majors right and we're you know hoping to list a lot more assets real soon 03:10.803 --> 03:13.168 [SPEAKER_02]: I love it, and we'd love to hear a little bit about it. 03:13.208 --> 03:18.860 [SPEAKER_02]: So let's just take a, they take a little step backwards here and talk specifically about harmonic. 03:18.880 --> 03:22.106 [SPEAKER_02]: I know you're the co-founder, I know you've been building it, you got a big vision. 03:22.147 --> 03:26.736 [SPEAKER_02]: So just for the entire audience, what's harmonic, what are you building, and what's the big vision? 03:27.526 --> 03:40.062 [SPEAKER_00]: Yeah, so think of Arbana because like an open transparent competitive block building marketplace for validers, that aims to be maximally configurable and really just have maximum optionality for validers. 03:40.102 --> 03:40.222 [SPEAKER_00]: Right. 03:40.242 --> 03:42.970 [SPEAKER_00]: So you have a lot of different validers on the network. 03:42.950 --> 03:47.738 [SPEAKER_00]: All those validators, they all have different preferences, they want different things, right? 03:48.218 --> 03:56.612 [SPEAKER_00]: Like, we are, we are very, you know, where, and we acknowledge the fact that every validator they want something different, right? 03:56.832 --> 03:59.617 [SPEAKER_00]: It is a decentralized network for overall, right? 03:59.637 --> 04:01.680 [SPEAKER_00]: And we really do believe in decentralization, right? 04:01.760 --> 04:02.241 [SPEAKER_00]: So, 04:02.221 --> 04:07.826 [SPEAKER_00]: You know, knowing there's, you know, hundreds of different validators on the network, we have to build a solution, right? 04:07.926 --> 04:10.168 [SPEAKER_00]: As the folks building this product that works for all of them. 04:10.308 --> 04:10.468 [SPEAKER_00]: Right. 04:10.488 --> 04:13.892 [SPEAKER_00]: And that means maximum option, all the maximum configuration, right? 04:14.232 --> 04:17.455 [SPEAKER_00]: And it also means creating maximum competition, right? 04:17.495 --> 04:20.057 [SPEAKER_00]: Because, and this kind of the second pillar, right? 04:20.077 --> 04:22.099 [SPEAKER_00]: We see competition as extremely important, right? 04:22.119 --> 04:27.163 [SPEAKER_00]: Like, we don't really want to see a world where block building is monopolized by anyone actor, right? 04:27.183 --> 04:32.228 [SPEAKER_00]: It's like we want to lower the bar to into the space 04:32.208 --> 04:36.091 [SPEAKER_00]: And then we're very aware that the incentives like, or what they are, right? 04:36.111 --> 04:39.214 [SPEAKER_00]: So like acknowledging that the incentives are what they are, right? 04:39.675 --> 04:43.778 [SPEAKER_00]: Let's make an open competitive marketplace that is as transparent as possible. 04:44.319 --> 04:49.343 [SPEAKER_00]: And aims to align incentives between everyone involved, users, apps, validators, the best that we possibly can. 04:51.345 --> 04:55.389 [SPEAKER_02]: So what made you focus on this part of the Solana ecosystem? 04:55.409 --> 04:55.889 [SPEAKER_02]: Why Solana? 04:55.909 --> 04:57.130 [SPEAKER_00]: Yeah. 04:57.170 --> 05:00.173 [SPEAKER_00]: So we'll say, why Solana or why Black building? 05:00.946 --> 05:02.578 [SPEAKER_00]: Uh, why block building? 05:02.658 --> 05:03.544 [SPEAKER_00]: Why block building? 05:03.565 --> 05:03.726 [SPEAKER_00]: Yes. 05:03.907 --> 05:04.048 [SPEAKER_00]: Okay. 05:04.068 --> 05:04.270 [SPEAKER_00]: Yes. 05:04.290 --> 05:05.337 [SPEAKER_00]: So soon. 05:06.110 --> 05:07.672 [SPEAKER_00]: It's a very typical problem for one, right? 05:07.712 --> 05:11.136 [SPEAKER_00]: So it's not a very crowded space by any means, right? 05:11.156 --> 05:15.582 [SPEAKER_00]: And it's a very uncounted space, because it's an extremely hard technical problem, right? 05:15.602 --> 05:17.644 [SPEAKER_00]: Like you need insanely good quantitative skills. 05:17.664 --> 05:22.630 [SPEAKER_00]: You need extremely good high performance computing skills across the board to efficiently tackle this problem. 05:22.650 --> 05:24.673 [SPEAKER_00]: Like there's so much spam on chain, right? 05:24.713 --> 05:26.455 [SPEAKER_00]: Like ordering is a really hard problem. 05:26.475 --> 05:29.899 [SPEAKER_00]: Like ordering is an NP-hard problem, right? 05:29.959 --> 05:33.163 [SPEAKER_00]: I mean, that just like an algorithm, academic speak just means it's, 05:33.143 --> 05:43.832 [SPEAKER_00]: really really hard to solve efficiently and then like a small amount of time as the number of transactions like you know pros and gross and gross and gross and gross the problem just gets like a harder at a very fast rate. 05:45.434 --> 06:01.868 [SPEAKER_00]: So if we were probably one of the only games on Solana quick like Tapahless problem right because we have this HFT, HPC like quantitative type backgrounds right incredibly hard problem this is all right like as a much harder than almost any of the other problems 06:01.848 --> 06:07.776 [SPEAKER_00]: And then the other thing is, right, we're really big leaders in competition and the competition can push the ecosystem forward, right? 06:08.197 --> 06:11.181 [SPEAKER_00]: And if we're being honest, the black building had no competition, right? 06:11.201 --> 06:14.846 [SPEAKER_00]: It was on track to be completely monopolized by a single actor, right? 06:15.167 --> 06:17.931 [SPEAKER_00]: Nothing against that actor at all, right? 06:18.251 --> 06:24.520 [SPEAKER_00]: But we think that's where the chain to thrive, it needs competition on every single vertical, and that includes black building, right? 06:24.540 --> 06:29.307 [SPEAKER_00]: So, you know, we're here, we're bringing competition hard, the black building, and you know, we're... 06:29.287 --> 06:32.692 [SPEAKER_00]: trying to make this network better for it all, and there's users and apps like. 06:35.375 --> 06:38.480 [SPEAKER_01]: I saw an update and maybe you're the best person to explain this. 06:38.720 --> 06:39.521 [SPEAKER_01]: It's about this. 06:39.541 --> 06:40.462 [SPEAKER_01]: It was by Solana, right? 06:40.482 --> 06:44.929 [SPEAKER_01]: I talked about the fair block construction guarantee that the protocol level is called constellation. 06:44.969 --> 06:46.791 [SPEAKER_01]: It's playing that to us, Normie. 06:46.831 --> 06:47.252 [SPEAKER_01]: What is that? 06:48.113 --> 06:50.416 [SPEAKER_01]: How do you guys fit in that picture? 06:50.977 --> 06:51.558 [SPEAKER_00]: Yes, sure. 06:51.598 --> 06:58.247 [SPEAKER_01]: So have you heard of multiple concurrent proposals before? 06:58.345 --> 07:00.389 [SPEAKER_00]: Yeah, so that's basically what constellation is. 07:00.410 --> 07:13.278 [SPEAKER_00]: It's an MCP multiple concurrent proposals proposition for Ritten Largely by the Ons of Research team, right, and is the proposals out, right? 07:13.318 --> 07:17.046 [SPEAKER_00]: It's probably going to have some revisions now currently in debate, some months ago. 07:17.347 --> 07:20.652 [SPEAKER_00]: all the eco teams and even within the eco teams, right? 07:21.073 --> 07:24.177 [SPEAKER_00]: But yeah, it's cool to finally see an MCP proposal out there. 07:24.197 --> 07:32.750 [SPEAKER_00]: You know, I had the opportunity to re-do the whole paper and to look at it pretty closely, so pretty familiar with how it works and the premise of what the proposal is. 07:33.351 --> 07:35.254 [SPEAKER_00]: Essentially with MCP, right? 07:35.234 --> 07:39.800 [SPEAKER_00]: There's one property that Solana needs that it does not have enough of, right? 07:39.820 --> 07:41.983 [SPEAKER_00]: That property is sensor-ship resistance, right? 07:42.023 --> 07:49.734 [SPEAKER_00]: So the question and the kind of the problem that MCP aims to solve is let's bring more sensor-ship resistance to the chain, right? 07:49.814 --> 07:57.224 [SPEAKER_00]: And MCP does that by instead of having one proposal who you have to go through, you can only send transactions to this one proposal. 07:57.585 --> 08:00.769 [SPEAKER_00]: You can send transactions up to 16 proposals. 08:00.749 --> 08:04.537 [SPEAKER_00]: and you can land through any of those 16 preposers, right? 08:04.717 --> 08:10.369 [SPEAKER_00]: So it sounds good on tables, like, oh, you know, censorship resistance, you can send to like all these guys and like, when you look through it, right? 08:10.409 --> 08:13.155 [SPEAKER_00]: Like, I think, it's, it's, it's, 08:13.793 --> 08:43.783 [SPEAKER_00]: there are some trade-offs right now there's trade-offs with everything that's not necessarily a bad thing right censorship resistance is really been important so like the question in our eyes is like is there a better easier way to achieve censorship resistance either by revising the existing constellation protocol which you know on this working on right now other teams like us you know an FD are providing a lot of feedback and then uh or there's another way that we're really bullish on to achieve much better censorship 08:43.763 --> 08:47.286 [SPEAKER_00]: which is just lowering sloth times a lot to like 20 mil seconds 50 mil. 08:47.306 --> 08:50.188 [SPEAKER_00]: So this is much simpler, way easier to, way easier to implement. 08:50.569 --> 08:55.793 [SPEAKER_00]: We'll consume probably less bandwidth and you know, have way less like network hops, right? 08:56.274 --> 09:04.060 [SPEAKER_00]: And then it's also when you look at like the fine print of the protocol like the constellation is selective censorship resistance, which is not like that's kind of the agony. 09:04.100 --> 09:06.983 [SPEAKER_00]: It's not like full-sensorship resistance, right? 09:07.123 --> 09:13.348 [SPEAKER_00]: Like there's still 09:13.328 --> 09:16.613 [SPEAKER_00]: you know make make life harder for for people channel interactions. 09:16.633 --> 09:27.488 [SPEAKER_00]: For example the leader can actually delay your transactions by up to 300 milliseconds and they can actually put competing transactions transactions that were sent later or paying less fee. 09:27.508 --> 09:35.640 [SPEAKER_00]: They can still put those before those higher fee transactions through through this thing that we're calling internally like 09:35.620 --> 09:37.622 [SPEAKER_00]: cross cycle MF, right? 09:37.742 --> 09:40.084 [SPEAKER_00]: So constellation introduces this thing called a cycle. 09:40.204 --> 09:41.145 [SPEAKER_00]: It's 50 milliseconds. 09:41.205 --> 09:46.210 [SPEAKER_00]: Transactions are priority ordered within a cycle, but the leader controls six cycles in a row. 09:46.631 --> 09:50.755 [SPEAKER_00]: So there are cross cycle MF opportunities that the leader can do. 09:50.795 --> 09:57.121 [SPEAKER_00]: And if the leader is cooperating with a proposer, proposer's the leaders are different things in this protocol. 09:57.541 --> 09:59.904 [SPEAKER_00]: There's a lot of types of MF that they can extract together. 10:00.124 --> 10:02.286 [SPEAKER_00]: And the leader proposer, 10:02.266 --> 10:04.289 [SPEAKER_00]: combination is actually very, very feasible. 10:04.329 --> 10:13.622 [SPEAKER_00]: So like I talked to on a little bit about some of this stuff and sounds like they're, you know, drafting up some revisions to the original constellation protocol just based on a lot of feedback. 10:13.983 --> 10:20.933 [SPEAKER_00]: In fact, we gave them, we gave them some feedback already to switch the tiebreaker from signature to a white and signature. 10:20.913 --> 10:31.228 [SPEAKER_00]: won't go to too much detail on what that is, but it basically just like before the white signature, you can basically have a bunch of GPUs and grind signatures to like try and get a good placement. 10:31.248 --> 10:39.401 [SPEAKER_00]: But with the white and signature, each signature changes the ordering in a way that's not predictable to any of the actors that are all submitting transactions. 10:39.521 --> 10:40.603 [SPEAKER_00]: So that's pretty desirable. 10:40.623 --> 10:44.288 [SPEAKER_00]: We already got one suggestion in there that I think is going to reduce spam a lot. 10:44.268 --> 10:59.914 [SPEAKER_00]: It's like, we'll see, I think the proposal as is, it's a very good proposal, huge congratulations to all the ons of researchers and everyone else involved, but I do think it needs some work, and then the other thing that I... 10:59.894 --> 11:18.173 [SPEAKER_00]: You don't want to see as a full comparison to, is like a full comparison to like all alternative proposals like 20 miles, like in 50 miles can slot times, because those could arguably make sense for ship resistance even better than it is in constellation with way less of the complexity, right? 11:18.153 --> 11:24.001 [SPEAKER_00]: Another thing that I wanted to see in the paper that's not there is where's the empirical motivation. 11:24.341 --> 11:29.809 [SPEAKER_00]: Like any detailed research paper, you know, is going to have some section on empirical motivation. 11:29.929 --> 11:39.042 [SPEAKER_00]: The paper claims that like single-purpose systems will basically become next extractive and they'll devolve into like max extract PBS. 11:39.222 --> 11:41.064 [SPEAKER_00]: There's like two sentences on economics. 11:41.625 --> 11:45.190 [SPEAKER_00]: It doesn't really engage with any of the counter-arguments at all, right? 11:45.170 --> 11:56.551 [SPEAKER_00]: competition can actually make life better for users, but the paper just probably claims that it makes life worse without any evidence or reasoning or data whatsoever, right? 11:56.952 --> 12:00.478 [SPEAKER_00]: So I I'd love to see a longer empirical motivation section. 12:00.458 --> 12:03.684 [SPEAKER_00]: on like one of the future iterations of the paper. 12:03.724 --> 12:07.210 [SPEAKER_00]: There's also vast implications for valid economics, right? 12:07.731 --> 12:13.621 [SPEAKER_00]: It mentions all these fees, but the paper doesn't actually really specify like what fees are burned, what fees go to validators. 12:14.002 --> 12:18.450 [SPEAKER_00]: There's one sentence that there's one sentence. 12:18.430 --> 12:27.382 [SPEAKER_00]: that basically mentions that fees are socialized, that they go to all validators based on state weight, but I think the plan is also to link burn a lot of the fees, right? 12:27.402 --> 12:40.600 [SPEAKER_00]: So this is going to vastly impact validator economics, and until they specify where these fees go, I'm kind of going to assume this means validators are going to earn less fees, because there's been a lot of talk internally amongst the ecosystems about just burning a lot of these fees. 12:40.660 --> 12:44.725 [SPEAKER_00]: So I mean, look, I'm going to guess validators are probably going to make less money. 12:44.705 --> 12:45.707 [SPEAKER_00]: under this proposal. 12:45.767 --> 12:49.973 [SPEAKER_00]: Now, I love for them to actually specify all the stuff on the paper, and we can reason about it a lot better. 12:50.434 --> 12:56.544 [SPEAKER_00]: The other thing I don't really like are the socialized fees, which we're kind of just calling fee communism internally. 12:57.105 --> 13:06.680 [SPEAKER_00]: Basically, right now, there's a lot of incentive to invest in better infrastructure, better CPUs, better network cables, better hardware, like across the board, better disks, right? 13:07.041 --> 13:07.962 [SPEAKER_00]: But 13:07.942 --> 13:23.318 [SPEAKER_00]: In the current constellation proposal, every single validator pays the price of a single bad potato in the network, because if you have a single bad potato in the network, they slow down the whole network and since fees are socialized, every single validator pays the price for it. 13:23.298 --> 13:24.419 [SPEAKER_00]: The flip side is also true. 13:24.439 --> 13:31.847 [SPEAKER_00]: If I have the best hardware on the network, everyone else benefits from my good hardware, but I'm spending the money on that good hardware and I don't get rewarded for that, right? 13:32.107 --> 13:36.992 [SPEAKER_00]: Today, we spend a ton of money on our valve ender hardware and we do it because that increases the rewards. 13:37.052 --> 13:39.154 [SPEAKER_00]: That increases the performance of our validers, right? 13:39.174 --> 13:42.818 [SPEAKER_00]: And if we do not reward validators, proportional to their performance, right? 13:43.018 --> 13:49.204 [SPEAKER_00]: Valorers are not going to be incentivized to invest in good hardware and we expect performance of the network to actually degrade. 13:49.876 --> 13:55.802 [SPEAKER_00]: So, let me just, all sorts of other things, but those are like a few of the things that popped out. 13:56.163 --> 13:57.965 [SPEAKER_02]: Let me break this down a little bit here. 13:58.005 --> 14:01.949 [SPEAKER_02]: So why different groups should care about what harmonic is building. 14:02.009 --> 14:13.462 [SPEAKER_02]: So let's talk about why validators care, and then you kind of just talked about that a little bit, but why validators should care, why devs in their dev team should care, and then why every day users. 14:13.722 --> 14:15.063 [SPEAKER_02]: So just kind of three points out. 14:15.103 --> 14:18.347 [SPEAKER_02]: Why validators, why devs, and why every day users. 14:18.412 --> 14:36.775 [SPEAKER_00]: I've got some moving from constellation and inter harmonic now and then you also just play my right like if constellation passes right we have to rewrite all the harmonic right so like that's it you know obviously it's obvious thing to to call out there right but you know we're all I think we're always always supportive of you know 14:36.755 --> 14:38.718 [SPEAKER_00]: something that makes the network better, right? 14:38.738 --> 14:42.384 [SPEAKER_00]: So, again, for us, like, on that point, right, I think we want a way of consolation. 14:42.404 --> 14:45.389 [SPEAKER_00]: We want to weigh all alternatives, like, 20, 50 miles a second, so lots of times. 14:45.469 --> 14:53.802 [SPEAKER_00]: And yeah, I mean, we really just want to, either of these would require a rewrite of harmonic, by the way, but we do, you know, want to make censorship resistance better on changing, right? 14:53.842 --> 14:57.027 [SPEAKER_00]: So, now, leading into harmonic, harmonic solves different problems, right? 14:57.047 --> 15:00.012 [SPEAKER_00]: So, harmonic solves other problems. 14:59.992 --> 15:06.239 [SPEAKER_00]: So, I mean, first of all, right from validators, we have a ton of demand across the board for a variety of different types of scheduling strategies, right? 15:06.299 --> 15:14.248 [SPEAKER_00]: So, like SFPP is now requiring you to either implement like FIFO or FIFI builds up an FBA, harmonics supports both of those, right? 15:14.489 --> 15:19.635 [SPEAKER_00]: We also have all sorts of other strategies that like validators can choose from, right? 15:19.655 --> 15:22.658 [SPEAKER_00]: So, like we give, we give validators, um... 15:22.638 --> 15:26.843 [SPEAKER_00]: maximum optionality in terms of the strategy they want, and this can mean all sorts of things, right? 15:26.864 --> 15:29.167 [SPEAKER_00]: It can mean high rewards for them, right? 15:29.187 --> 15:33.112 [SPEAKER_00]: Like, it can mean compliance with, like, what, what, like any kind of statement wants, right? 15:33.172 --> 15:42.604 [SPEAKER_00]: You know, we're very, like, the compliance with, like, STP and, like, everyone else is, like, it's super important for us, but, like, obviously, want what is best for the network rights, like, we're getting validators, like, 15:42.584 --> 15:43.706 [SPEAKER_00]: all these options, right? 15:43.726 --> 15:49.957 [SPEAKER_00]: And we're even recommending the same options that SFDP is recommending to them, right? 15:49.977 --> 15:53.183 [SPEAKER_00]: It's like for validators, it means compliance, which is really, really important, right? 15:53.203 --> 15:56.088 [SPEAKER_00]: It means higher rewards, right? 15:56.489 --> 15:58.332 [SPEAKER_00]: Because, like, again, like given 15:58.312 --> 16:06.184 [SPEAKER_00]: our background that we just explained, we can create black building strategies that just, you know, will increase rewards on a crazy scale, right? 16:06.665 --> 16:19.464 [SPEAKER_00]: And, you know, one thing that's tough, even with 50 mills like an FDA or FIFA, is there's still a lot of the invalid spam, and all sorts of spam and doses, and you still need to be able to ingest transactions extremely fast and filter through the spam extremely fast, right? 16:19.484 --> 16:26.715 [SPEAKER_00]: Like there's so many transactions hitting the quick servers of our 16:26.695 --> 16:35.311 [SPEAKER_00]: being able to pack these extremely fast and being able to filter through transactions really fast ultimately does translate to higher rewards. 16:35.331 --> 16:37.776 [SPEAKER_00]: We've also with harmonically removed all the rate limits. 16:37.796 --> 16:42.324 [SPEAKER_00]: This means higher rewards for validators, but it's also actually really good for allocations and users. 16:42.344 --> 16:43.827 [SPEAKER_00]: It's like one of the original 16:43.807 --> 16:48.274 [SPEAKER_00]: motivations of harmonic was we kept running into rate limits on Gito, right? 16:48.614 --> 16:52.640 [SPEAKER_00]: Like Gito has really, really strict rate limits and they're like completely arbitrary, right? 16:53.021 --> 17:00.391 [SPEAKER_00]: And then there's also state-weighted QOS rate limits, which the ECOVA client has, that we've also removed in harmonic, right? 17:00.411 --> 17:09.004 [SPEAKER_00]: Removing all these rate limits basically means that, you know, we can ingest way more transactions that creates better landing experience for applications and users, a better and more predicted. 17:08.984 --> 17:12.329 [SPEAKER_00]: But landing experience, it also means rewards are higher for validators, right? 17:12.349 --> 17:23.407 [SPEAKER_00]: So this is an easy win-win scenario between validators, allocations and users, which is getting rid of these rate limits and having a much more performance, you know, filtering stock for invalid transaction spam. 17:24.449 --> 17:26.713 [SPEAKER_00]: So yeah, that was a massive improvement there. 17:26.733 --> 17:29.878 [SPEAKER_00]: If I kind of answer all the questions there, I mean, there's, you know, 17:29.858 --> 17:36.633 [SPEAKER_00]: A lot more tooling that we're building for apps and users, we want to make life for apps and users as easy as possible. 17:36.673 --> 17:41.303 [SPEAKER_00]: We want to give them more visibility on what actually happens to their transaction after they send it. 17:41.323 --> 17:44.189 [SPEAKER_00]: We want to give them better feedback on how to land their transactions better. 17:44.750 --> 17:47.496 [SPEAKER_00]: With harmonic we're building out all this application. 17:47.476 --> 17:54.305 [SPEAKER_00]: user-focused tooling that we can hopefully use to make UX better for the chain across the board. 17:54.345 --> 18:06.081 [SPEAKER_00]: And for developers, it's relevant for the same reasons, like incendiary transactions, like incident directly to harmonic or even if they're sending over the normal TPU, like, you're much more likely to land in a much, much, much timely manner, right? 18:06.101 --> 18:08.404 [SPEAKER_00]: Because we can ingest just, 18:08.384 --> 18:17.251 [SPEAKER_00]: So much to what we also provide them options like reverb protection, that's not, and like these options are not available on any of the default lines. 18:17.271 --> 18:22.326 [SPEAKER_00]: So we build even more features that the applications want as well and reverb protection is a big one. 18:23.268 --> 18:26.434 [SPEAKER_01]: What's the, as a retail user for Solana? 18:26.995 --> 18:27.296 [SPEAKER_01]: Yeah. 18:27.316 --> 18:30.061 [SPEAKER_01]: The benefits seemed largely for institutions, right? 18:30.181 --> 18:35.071 [SPEAKER_01]: They required this, you know, faster, fair, tighter pricing, less slippage, less bought advantage. 18:35.893 --> 18:42.205 [SPEAKER_01]: But as a just a regular retail user in Solana, if you're pretty quick already, where does the retail user see the benefits? 18:42.265 --> 18:44.630 [SPEAKER_01]: Is it just in pricing and more speed? 18:44.690 --> 18:45.411 [SPEAKER_01]: Yeah. 18:45.391 --> 19:02.454 [SPEAKER_00]: Yeah, so I mean, obviously there will be a lot of benefits that are like downstream of all of this that are also going to like, you know, affect users, right, like a simple one is since we removed the great moments, we can lay in more or collect dates on chain, letting more or collect dates on chain, these more communicers can quote tighter, that translates into better spreads across the board, right. 19:03.636 --> 19:07.101 [SPEAKER_00]: Revert protection is another really useful one for users. 19:07.802 --> 19:07.922 [UNKNOWN]: Right. 19:07.902 --> 19:12.987 [SPEAKER_00]: Well, sometimes users long send a transaction and it fails and they waste the fee, right? 19:13.547 --> 19:18.331 [SPEAKER_00]: Reverb production means that the transaction, uh, if it fails, it'll just revert and never land on chain. 19:18.892 --> 19:20.313 [SPEAKER_00]: So they will waste the fee. 19:20.393 --> 19:21.654 [SPEAKER_00]: This one's all very good for users. 19:21.694 --> 19:32.965 [SPEAKER_00]: And look, if you think it's fast now, like you might actually notice, you know, over the next few months, since, you know, it's not, not all of these, the work we're doing because the work they ever want is doing, that UX is actually going to get better and better, better. 19:32.985 --> 19:37.909 [SPEAKER_00]: Like it's going to get snappy and 19:37.889 --> 19:45.861 [SPEAKER_00]: But yeah, and like we, we need salons to be as seamless as like these centralized Central other change, you eyes are basically, right? 19:46.282 --> 19:49.026 [SPEAKER_00]: Like that's that's how slim the seamless we need to be, right? 19:49.046 --> 19:51.449 [SPEAKER_00]: They, the centralized exchanges are the real competitors here. 19:51.590 --> 19:54.334 [SPEAKER_00]: Centralized finance are the real competitors here, right? 19:54.354 --> 19:58.480 [SPEAKER_00]: So like that's that's really the bar and until we're beating them, we're in fear here, right? 19:58.848 --> 20:05.213 [SPEAKER_02]: Do you think if a chapter improvements like this are part of what will bring larger pools of money or capital on chain? 20:05.714 --> 20:07.375 [SPEAKER_00]: Yeah, yeah, yeah, I mean, like for sure, right? 20:07.395 --> 20:10.958 [SPEAKER_00]: Like when, you know, when the network is not congested, right? 20:10.978 --> 20:12.840 [SPEAKER_00]: And like when there's way more throughput, right? 20:12.860 --> 20:15.142 [SPEAKER_00]: And there's way lower latency, right? 20:15.302 --> 20:22.748 [SPEAKER_00]: I mean, that that I think decreases risk across the board for everyone, which does obviously allow, you know, less risk for larger liquidity, right? 20:22.768 --> 20:28.753 [SPEAKER_00]: Because as you move more and more and more 20:29.172 --> 20:35.225 [SPEAKER_02]: So why do you think Solana is becoming a more serious destination for bigger capital and more sophisticated in use cases? 20:35.325 --> 20:42.600 [SPEAKER_02]: Because a lot, we've obviously the speed in the ease of the views and we're big fans of Solana over here as well. 20:43.141 --> 20:48.312 [SPEAKER_02]: But we'd love to just kind of have this snapshot of Solana and that ecosystem from you. 20:49.304 --> 21:08.342 [SPEAKER_00]: yeah so i mean it's like a few reasons right like obviously like one is that the tech has gotten a lot better right like better tech is just the foundation though right it's something that was necessary but not sufficient right like in parallel i'm sure you've seen this huge push by like salana foundation right i'm like a lot of other folks in uh 21:08.322 --> 21:16.075 [SPEAKER_00]: Sloan, like, on the institutional stuff, we're seeing more payments and stuff come on chain. 21:16.115 --> 21:22.085 [SPEAKER_00]: We're seeing more and more big actors come on chain and start deploying their capital into DeFi. 21:22.406 --> 21:29.858 [SPEAKER_00]: We're seeing a lot more focus on the board, not just on the tech, but also on institutional relations and institutional outreach. 21:30.339 --> 21:30.499 [SPEAKER_00]: Even 21:30.479 --> 21:32.383 [SPEAKER_00]: know, we're a decentralized network, right? 21:32.403 --> 21:37.512 [SPEAKER_00]: Like, you know, people from the ecosystem still need to go out, talk to these institutions, right? 21:37.532 --> 21:40.939 [SPEAKER_00]: Because as you know, institutions are extremely risk of earths, right? 21:40.959 --> 21:42.842 [SPEAKER_00]: And like, we know this from harmonic first time, right? 21:43.263 --> 21:44.826 [SPEAKER_00]: Institutions love harmonic, right? 21:44.846 --> 21:46.810 [SPEAKER_00]: Because harmonic gives them better yield, right? 21:46.830 --> 21:47.391 [SPEAKER_00]: And it's, 21:47.371 --> 22:16.197 [SPEAKER_00]: like we're giving them more visibility into what's going on right and this kind of visibility is exactly what the institutions want and this yield is obviously what the institutions want right it's like we have yield but then like as we were talking about before right like it becomes a factor of like risk and there's risk on many levels right there's compliance and regulatory risk right and right now it's looking like we're finally going to start to get more regulatory clarity right like I think the you know environment especially in the US in the last few 22:16.177 --> 22:19.264 [SPEAKER_00]: has definitely been more friendly to defy. 22:19.284 --> 22:21.649 [SPEAKER_00]: It's been more friendly to crypto in general, right? 22:21.749 --> 22:32.212 [SPEAKER_00]: And that more friendly regulatory environment has definitely made institutions a lot less of worse to deploying their capital on chain and like participating more in defy and actually using the chain. 22:32.232 --> 22:35.198 [SPEAKER_00]: So that's like one huge win and then like, you know, 22:35.178 --> 22:58.452 [SPEAKER_00]: other main thing is like we've seen a lot of teams on the ecosystem and you know also like Solana Foundation included you know present a much more sophisticated and unified front to these institutions right like the work that Solana Foundation is doing with the institutions is pretty crazy right like they've they've obviously put a ton of work into onboarding institutions even they're launching like entire portals and stuff 22:58.432 --> 23:17.622 [SPEAKER_00]: for like institutional relations and institutional trading right and across the board we're gonna see more ecosystem teams do that as well right because like I mean the institutions have all the capital right we might have the yield we might have the tech but we're new kids on the block right like we have progress to make on the regulatory front and we have progress to make on the institutional relations front right like 23:17.602 --> 23:23.755 [SPEAKER_00]: If institutions don't know what the hell is going on on TFI, where the yields coming from, they're not going to use any of it, right? 23:23.775 --> 23:26.481 [SPEAKER_00]: If it's not compliant, they're not going to use any of it. 23:26.781 --> 23:26.982 [SPEAKER_00]: Right. 23:27.483 --> 23:35.239 [SPEAKER_00]: And so yeah, it's like the combination of better installations and more regulatory clarity would be my take on it. 23:35.489 --> 23:41.638 [SPEAKER_01]: I'm going to go on a tangent here and you seem like the best person to answer this because you're deep in the text deck. 23:42.459 --> 23:47.307 [SPEAKER_01]: How is AI impacting your work at harmonic and the only piece of land overall? 23:47.367 --> 23:48.128 [SPEAKER_01]: You know, are they leading? 23:48.268 --> 23:50.211 [SPEAKER_01]: Is there a chain that's leading in that direction? 23:50.251 --> 23:51.353 [SPEAKER_01]: And we know Tao, right? 23:51.593 --> 23:55.960 [SPEAKER_01]: Yeah, but who's implementing it in a correct way to improve the chain overall? 23:55.980 --> 24:04.973 [SPEAKER_00]: Yes, so are we talking about AI in like built on Solano or like just like using AI to help build things on Solano? 24:05.122 --> 24:11.352 [SPEAKER_01]: Yeah, it's a more general question because I want to make sure that I'm asking the most impactful question. 24:11.372 --> 24:13.175 [SPEAKER_01]: Maybe you have to answer that. 24:13.195 --> 24:13.395 [SPEAKER_00]: Yeah. 24:13.415 --> 24:13.616 [SPEAKER_00]: Okay. 24:13.636 --> 24:17.602 [SPEAKER_00]: So like firstly, like AI is obviously impacted the life of every developer on the planet. 24:17.622 --> 24:17.742 [SPEAKER_00]: Right? 24:17.763 --> 24:21.028 [SPEAKER_00]: Like if developers not using AI to help them code, they're already behind. 24:21.148 --> 24:21.609 [SPEAKER_00]: Right? 24:21.629 --> 24:30.403 [SPEAKER_00]: Now like I do warn any developers out there letting AI do literally everything because I definitely do not trust AI enough to do like literally everything. 24:30.383 --> 24:38.533 [SPEAKER_00]: That's said, like, my army of clods are, like, chatshebyTs, like, they're good, like, low-level engineers, but they need very, very tight supervision, right? 24:38.554 --> 24:46.724 [SPEAKER_00]: So, like, I would say, like, use the AI as an engineer to amplify your output and to amplify your throughput and not use the AI to replace you, right? 24:46.764 --> 24:48.687 [SPEAKER_00]: If you do one, you'll be way more productive. 24:48.787 --> 24:54.013 [SPEAKER_00]: If you do two, you're gonna end up pushing code with, like, critical vulnerabilities or, like, issues, right? 24:54.033 --> 24:59.901 [SPEAKER_00]: Cause, like, look, I mean, but AI's not good enough yet to just autonomously do stuff on its own and push, like, good. 24:59.881 --> 25:28.386 [SPEAKER_00]: see it good secure code maybe one day will be but right now it's as good as like a low level like in turn our engineer and the nice thing is it's like a scalable low level internal engineer right so like a bunch of low level like grunt work to do I can now just like delegate that to like you know AI agents and stuff right which like like it's really awful right like it's improved our throughput a lot you know temporal harmonic every other team you know we expect all of our competitors as well are also starting to adopt these tools and using these tools to 25:28.366 --> 25:31.533 [SPEAKER_00]: amplify their own throughput as well. 25:32.034 --> 25:35.061 [SPEAKER_00]: AI infrastructure like build on Slon is also interesting, right? 25:35.682 --> 25:38.328 [SPEAKER_00]: You mentioned Tau, I don't know too much about Tau or I don't know. 25:38.348 --> 25:41.836 [SPEAKER_00]: There's some marketing that like their AI infrastructure and stuff. 25:41.856 --> 25:44.842 [SPEAKER_00]: I don't really look to anything beyond the marketing, right? 25:44.862 --> 25:47.067 [SPEAKER_00]: So I can't comment much on that. 25:47.047 --> 25:54.674 [SPEAKER_00]: But yeah, look, I will say I think a power and energy, I mean, there's a few things that are going to be with limiting product or an AI. 25:54.814 --> 25:56.495 [SPEAKER_00]: One is obviously data, right? 25:57.256 --> 26:03.662 [SPEAKER_00]: We've just seen, as you put more data into these models, the models get better, and we've not seen that plateau yet, right? 26:03.682 --> 26:07.725 [SPEAKER_00]: So the folks are going to be in all these top AI labs, they are doing everything they can. 26:07.825 --> 26:11.749 [SPEAKER_00]: As much data as possible, data is really, really, really important, right? 26:11.769 --> 26:16.513 [SPEAKER_00]: In the LLMs, they're just getting better and better and better, 26:16.493 --> 26:18.516 [SPEAKER_00]: who knows, right? 26:18.737 --> 26:20.399 [SPEAKER_00]: The other thing that was powered, right? 26:20.559 --> 26:31.757 [SPEAKER_00]: The more that you have, the more power and the more, you know, the more compute power you need to like, you know, get through, get through a training epoch through like all the data, right? 26:31.777 --> 26:34.041 [SPEAKER_00]: Like just more data that you're training like the models on, right? 26:34.101 --> 26:36.004 [SPEAKER_00]: And that requires more and more and more power. 26:35.984 --> 26:42.200 [SPEAKER_00]: They also need to be able to experiment with tons of different models at once and do tons and tons of experiments in parallel. 26:42.541 --> 26:49.319 [SPEAKER_00]: I can't imagine the power bills and the GPU bills of these top AI labs like open AI and like anthropic AI. 26:49.339 --> 26:50.402 [SPEAKER_00]: They have to be insane. 26:50.502 --> 26:51.124 [SPEAKER_00]: I have 26:51.104 --> 27:10.895 [SPEAKER_00]: No idea how much they're spending on power right, but if they're raising like hundreds of billions of dollars right like I imagine infrastructure is going to be by far one of their biggest expenses right up there with the top talent That exists at these labs if not even higher right I'm not you like that'd be my guess of what the breakdown of expenses Kind of looks like at these AI labs you've got power energy infrastructure. 27:10.915 --> 27:15.222 [SPEAKER_00]: You've got talent and it's pretty much like those two things right 27:15.202 --> 27:22.073 [SPEAKER_00]: So with power also as a limiting factor right that you might actually see some deep-hand projects out there that like 27:22.390 --> 27:26.656 [SPEAKER_00]: allow people to build like decentralized power networks or even like decentralized power grids. 27:26.917 --> 27:28.179 [SPEAKER_00]: It's really really interesting at DEI. 27:28.199 --> 27:29.440 [SPEAKER_00]: I think there's a few projects like this. 27:29.480 --> 27:32.625 [SPEAKER_00]: Like I saw like fuse which claim to be like deep pain energy. 27:33.086 --> 27:44.242 [SPEAKER_00]: I'm not looked too much into these projects, but as you know power becomes a limiting factor in training AI and a limiting factor in making these AI models better and better, better, better, right? 27:44.262 --> 27:47.567 [SPEAKER_00]: Look, I think anything that makes power cheaper and more accessible is good, right? 27:47.587 --> 27:51.553 [SPEAKER_00]: And I think if you're doing that in a 27:52.495 --> 27:56.922 [SPEAKER_02]: You know, the power with the AI and the chatbot is funny. 27:57.022 --> 28:08.862 [SPEAKER_02]: What's in its seem, Altman, like a few weeks ago, a few months ago, that said, like, phrases, like, please and thank you have added up to, like, tens and millions of dollars of dollars spent. 28:09.142 --> 28:11.466 [SPEAKER_02]: Am I making that up, or was that actually said? 28:12.155 --> 28:15.858 [SPEAKER_00]: No, I mean, I don't know if he said that I have to check, but I do believe it, right? 28:15.918 --> 28:19.922 [SPEAKER_00]: Because look, I mean, it's really expensive to do inference on these models, right? 28:19.942 --> 28:30.091 [SPEAKER_00]: Like, you cannot put, like, cloud 4.6 or like KBT 5.4, like you can't put these models into your TPUs at home, right? 28:30.451 --> 28:34.875 [SPEAKER_00]: Just fitting this model, like requires a ton of hardware that nobody has at home. 28:35.256 --> 28:42.162 [SPEAKER_00]: And then even running the inference on this model that already exists is like a really, really expensive, right? 28:42.142 --> 28:43.945 [SPEAKER_00]: It's quite expensive. 28:44.125 --> 28:53.780 [SPEAKER_00]: So I would, yeah, I can't at least believe that, like, and that, like, enough thank yous and, like, please, is that, like, I would definitely believe that those those answers and those responses can add up to that amount of money. 28:53.801 --> 28:56.224 [SPEAKER_00]: Like, it is not cheap to run this infrastructure. 28:56.405 --> 29:04.217 [SPEAKER_00]: And, you know, the top AI models of today that, like, users are using, these are heavily subsidized by, like, these VC funds that are 29:04.197 --> 29:10.990 [SPEAKER_00]: a guarantee if you have a chat, you can see your cloud subscription and your power user always hitting your limits that they are losing money on you. 29:11.751 --> 29:17.061 [SPEAKER_00]: And they're losing money on you now so they have your business in the future and that they can increase prices on you in the future. 29:17.241 --> 29:20.367 [SPEAKER_00]: These, these, these, they have to turn a profit. 29:20.347 --> 29:22.169 [SPEAKER_00]: They have to turn a profit, right? 29:22.210 --> 29:25.254 [SPEAKER_00]: Like if they don't turn a profit, this whole bubble is just going to burst, right? 29:25.554 --> 29:32.123 [SPEAKER_00]: Now granted, the infrastructure is very useful and it has great implications for pretty much everyone all of the world, right? 29:32.143 --> 29:34.646 [SPEAKER_00]: Like the time has never been bigger, right? 29:34.726 --> 29:37.510 [SPEAKER_00]: Like the market is the entire world for these products, right? 29:37.530 --> 29:46.202 [SPEAKER_00]: So yeah, look, I think that, you know, they'll probably get more expensive, but like if they're useful enough to justify the cost, I do actually expect these labs can like turn a profit. 29:47.431 --> 29:50.796 [SPEAKER_02]: So let's put it a little bit back to Solana ecosystem. 29:50.856 --> 29:53.540 [SPEAKER_02]: Where do you think the biggest opportunities are for Solana? 29:53.560 --> 29:55.343 [SPEAKER_02]: Let's say, over the next year or two. 29:56.425 --> 30:01.993 [SPEAKER_00]: Yeah, so look, I envision Solana as becoming basically the retail chain, right? 30:02.013 --> 30:05.659 [SPEAKER_00]: Now, it'll also be really, really useful for institutions. 30:06.039 --> 30:11.888 [SPEAKER_00]: And we've been doing very good focusing on institutions, but I also want to see a focus on retail more, right? 30:12.048 --> 30:16.375 [SPEAKER_00]: The retail numbers on Solana have not been the best recently 30:16.355 --> 30:18.579 [SPEAKER_00]: Um, we can fix this through all sorts of ways, right? 30:18.599 --> 30:22.105 [SPEAKER_00]: It's like, one is like, we need better, we need more assets, right? 30:22.145 --> 30:23.968 [SPEAKER_00]: So we want to list more and more and more assets. 30:24.168 --> 30:25.610 [SPEAKER_00]: Humidify is going to be a big part of it. 30:26.011 --> 30:30.919 [SPEAKER_00]: We want, we want Solana to be the chain to trade anything, anywhere, any time. 30:31.160 --> 30:33.984 [SPEAKER_00]: That's like pretty much my vision for the chain over the next season too, right? 30:34.024 --> 30:36.649 [SPEAKER_00]: And I think we can get it there over the next season too, right? 30:36.669 --> 30:42.278 [SPEAKER_00]: Trade anything, anywhere, anytime, every asset, every asset class. 30:42.258 --> 30:42.579 [SPEAKER_00]: Right. 30:42.899 --> 30:44.723 [SPEAKER_00]: So right now we're really, really good on Seoul. 30:45.244 --> 30:46.947 [SPEAKER_00]: We have better spreads than the central exchange. 30:47.668 --> 30:49.371 [SPEAKER_00]: Just like why don't we do that for everything, right? 30:49.612 --> 30:51.054 [SPEAKER_00]: We can absolutely do that for everything. 30:51.094 --> 30:52.938 [SPEAKER_00]: Oh, this is going to take work, right? 30:53.178 --> 31:00.672 [SPEAKER_00]: But look, I mean, you're trading on a central exchange is like, you know, like if you're just like you have a normal retail account, and especially if you're in the US, 31:00.652 --> 31:04.539 [SPEAKER_00]: You're going to get clips by like a hundred bits fee like whatever you swap. 31:04.559 --> 31:05.420 [SPEAKER_00]: It's ridiculous, right? 31:05.440 --> 31:06.843 [SPEAKER_00]: Like you swap a hundred thousand dollars. 31:06.863 --> 31:08.325 [SPEAKER_00]: You just lost a thousand dollars. 31:08.706 --> 31:09.748 [SPEAKER_00]: You swap a million dollars. 31:09.788 --> 31:12.092 [SPEAKER_00]: You just lost like at least like ten thousand dollars. 31:12.112 --> 31:15.137 [SPEAKER_00]: Like that's not even counting the spreads on half of it, right? 31:15.417 --> 31:16.399 [SPEAKER_00]: So like 31:16.379 --> 31:25.456 [SPEAKER_00]: Clearly, there's a lot of money that can be saved and there's still enough port to go around for like defying like the market-makers capturing spreads, right, so look, let's take what we've done for Solano. 31:25.476 --> 31:37.459 [SPEAKER_00]: Let's scale it to every asset on every asset class, let's scale it to commodities, let's scale to 4x, let's scale to equities, we have x stocks down, right, let's scale it to perps, right, everything possible, right, we have something that works. 31:37.439 --> 31:43.891 [SPEAKER_00]: So like we improvement that Solana is a chain where spreads can be that tight and liquidity can be that deep. 31:44.312 --> 31:53.389 [SPEAKER_00]: So I expect liquidity to get a lot deeper and we're going to make damn sure this chain becomes the chain where you can trade anything anywhere anytime. 31:53.960 --> 31:54.582 [SPEAKER_02]: I love that. 31:54.722 --> 31:55.163 [SPEAKER_02]: I love that. 31:55.183 --> 31:56.948 [SPEAKER_02]: It sounds like a commercial for Solana. 31:57.008 --> 31:58.673 [SPEAKER_02]: So the Solana people listen up. 31:58.693 --> 32:01.982 [SPEAKER_02]: We got Ben, uh, reaching the truth right now. 32:02.744 --> 32:04.970 [SPEAKER_02]: I guess what my final question is here is looking ahead. 32:05.030 --> 32:09.943 [SPEAKER_02]: What's the long term vision for harmonic and what role do you want to play in Solana's future? 32:10.210 --> 32:17.562 [SPEAKER_00]: Yeah, so I mean, Harmonic is a product primarily for validators, perhaps, and users, right? 32:17.622 --> 32:26.698 [SPEAKER_00]: So really, really what we're going to be doing with Harmonic is, I mean, we're going to keep talking to our customers, and we're going to keep improving the product, right? 32:26.718 --> 32:31.446 [SPEAKER_00]: Like, I expect Harmonic to evolve based on app, user, and validator needs, right? 32:31.486 --> 32:35.813 [SPEAKER_00]: So the current focus for Harmonic is maximizing optionality, right? 32:35.793 --> 32:44.869 [SPEAKER_00]: We traditionally only had a few block pack and strategies and we're moving into like dozens of block packing strategies and each of those packing strategies is configurable. 32:45.089 --> 32:47.574 [SPEAKER_00]: And we're doing this just because this is what validators want. 32:47.814 --> 32:50.459 [SPEAKER_00]: Like validators have told us that they want more options. 32:50.499 --> 32:51.461 [SPEAKER_00]: They want more strategies. 32:51.521 --> 32:52.783 [SPEAKER_00]: They want more configurations. 32:53.284 --> 32:54.686 [SPEAKER_00]: They want more visibility. 32:54.726 --> 32:55.047 [SPEAKER_00]: Right. 32:55.067 --> 32:55.708 [SPEAKER_00]: So... 32:55.688 --> 33:14.490 [SPEAKER_00]: Currently we're really really focused on meeting validators needs and onboarding more validators to harmonics have their needs changed right now the other thing that we're paying really close attention to is The apps and the users right so in parallel we're also talking to like we have very good relationships with all the top applications like especially the trading apps on a on a salana 33:14.470 --> 33:16.893 [SPEAKER_00]: They have a lot of feature requests for us as well, right? 33:17.074 --> 33:20.879 [SPEAKER_00]: They want more visibility in what happens and they're transaction when they send it, right? 33:21.159 --> 33:23.823 [SPEAKER_00]: They want better planning rates, right? 33:23.843 --> 33:29.391 [SPEAKER_00]: So we're working with all the applications on tooling, they can help them improve all these things as well, right? 33:29.771 --> 33:31.334 [SPEAKER_00]: And look, we're very adaptable. 33:31.374 --> 33:34.478 [SPEAKER_00]: So those are the needs of the applications, users and validators change. 33:34.538 --> 33:36.681 [SPEAKER_00]: Our offerings are going to change as well. 33:38.011 --> 33:47.194 [SPEAKER_02]: Yeah awesome stuff so let's give it a little bit of an opportunity for everybody in audience if they want to learn more about you they want to learn more about harmonic how can they get in touch. 33:48.153 --> 34:17.451 [SPEAKER_00]: Yeah, so obviously we're all over Twitter right temporal underscore XYZ we're also like I think harmonic underscore GG I believe those would be the the names for our two main Twitter You can follow me on Twitter as well at hyponim's temporal website is temporal dot XYZ harmonic the website is harmonic dot GG and then humidify is humidify dot XYZ 34:17.431 --> 34:22.075 [SPEAKER_00]: Humidify, yeah, Twitter is where most things happen and like most things are announced. 34:22.095 --> 34:47.438 [SPEAKER_00]: So definitely stay in touch on Twitter And then I guess for like specific product offerings like Nizumi Nizumi is used up temporal that XYZ and then the docs on parmonic had to send both bundles and have on board As a valid or our docs dot harmonic dot GG and if anyone wants to email us or get on or get in touch We have hello at temporal dot XYZ 34:47.418 --> 35:00.388 [SPEAKER_02]: great stuff then and for the audience we'll get put all those links down below in this show description so you can find it easily and learn more about what they're building Ben really appreciate you jumping on the podcast today. 35:00.408 --> 35:01.892 [SPEAKER_00]: Amazing yes thanks for having those. 35:01.912 --> 35:02.473 [SPEAKER_00]: This is fun.