Is Lighter Ethereum's L2 Perp DEX? | Founder Vlad Novakovski

Ryan:
[0:02] Welcome to Bankless, where today we explore the frontier of Perp Dexas. We're here to help you become more bankless. You got Ryan and David on the podcast today. There's been a huge breakout this cycle. One of the themes has been decentralized Perps exchanges, led primarily by a application and a blockchain called Hyperliquid. But now some competition has entered the ring. And one in particular is moving up the charts in terms of open interest for Perps and volume. It's KotRI. It's called LIDR. It's now the second largest purpose exchange. They're only two weeks in to mainnet. So it's quite a feat. It's getting a lot of traction. The difference here is that LIDR is built on Ethereum. So the question is, is this Ethereum's hyperliquid? In this conversation, I think you'll find this is not just a battle of purpose exchanges. It's certainly that. But it's also a battle for Ethereum's vision of the world, which is L2s backed by Ethereum-grade security versus maybe some other visions of the world, a multi-chain vision of the world where every single app has its own new fresh layer one. A few things we discuss in this episode, the benefits of an L2.

Ryan:
[1:13] Can L2s really scale? Also, the massive flash crash on October 11th. And what were the causes? Vlad also gives letter grades to all of the major perps exchanges in terms of how they operated. We also discussed the LIDR roadmap and their upcoming token, including how to earn the airdrop and the future of LIDR and their plans to become a platform chain.

David:
[1:35] I really enjoyed talking to Vlad and just getting his perspective as like, why build a Layer 2 perp decks? And we didn't really specifically ask Vlad this question, but I don't think he's doing it because he believes in Ethereum or wants to do the layer two model. He's doing it because between ZK and between a layer two, he can build what he thinks is the fastest, more performant perp decks. And so it's not a question of ideology. It's a question of just can he build the most technically performant chain? Let's go ahead and get right into the conversation with Vlad from Leiter. Bankless Nation, we are here with Vlad Novikovsky, the founder and CEO of Leiter, a newly launched ZKL2 Perpdex on Ethereum, a ZKL2 Perpdex, that's fun to say,

David:
[2:19] which recently crossed $1 billion of TVL deposited all in USEC. Vlad, welcome to Bankless.

Vlad:
[2:25] Thanks, David. Thanks, Ryan.

David:
[2:26] Good to be here. I want to read out a tweet to get us started here because I enjoyed this tweet. Ryan enjoyed this tweet. This is a tweet from you. You said, being a layer one is a bug, not a feature. A layer one is just an Ethereum layer two without any of the security and virifiability parts. So Vlad, why did you build LIDAR as a ZK layer two?

Vlad:
[2:44] Yeah, well, I think that one, obviously it was a little bit tongue in cheek, but there are some very good ones as well, like Bitcoin is an obvious example. But I guess my point was that, right, when we were making this architectural choice, like, The hard part was inheriting the security layer of Ethereum while still having really good performance. And it's like, if you don't do that, then the way you can achieve security and verifiability is a much harder problem to kind of rebuild all of the functionality that Ethereum already has. So it's like, because a lot of people would say things like, oh, well, like after we built Lighter. You know, it's one thing in the beginning, maybe we could have made a different choice and we were very happy with the choice we made. Right. But like after we'd done all the hard work and made it work as an L2, some people were saying, oh, well, later I will be in L1. And it's like, yeah, all we have to do is just remove all the hard work we did and call it an L1. Like it's like like having like having the good performance and the verifiability and the security that that was always our goal. And we think we are, you know, well on our way to achieving it with the

Ryan:
[3:58] Structure we have. Can you get into that, Vlad? So we'll talk about performance maybe later, but the things that you get as an Ethereum layer to the security piece, the verifiability piece. What do you get by virtue of being an Ethereum layer to?

Vlad:
[4:11] Right, right. So the security piece is insured through something called an escape hatch that we have. And, you know, if you look at our white paper, it explains that in kind of technical detail. But also if you go to L2 beat, you know, it's where one of the L2s that has that. But basically what the escape hatch means is that if the L2 messes up or does stuff that is outside of the rules or if it gets hacked or, you know, kind of or let's say our own team somehow gets corrupted or whatever happens. If that happens, then, you know, within a certain time frame, if it's not fixed, customers can get their assets back through Ethereum. And so that's kind of, you know, a very strong way both to disincentivize bad actors, but also to ensure, you know, security. And so then, you know, and which clearly you don't have, if you're building your own L1, if that L1 is facing a problem,

Vlad:
[5:12] Like, that's it.

Vlad:
[5:13] I mean, it kind of lights out. Whereas here, obviously, we don't want that scenario to happen. But A, it's like strong and disincentivized. And B, if it did happen, customers can still get their assets back.

Ryan:
[5:24] So like, just so it's clear, what happens if a, maybe we'll not say Binance, but say something like a hyperliquid, let's say hyperliquid goes offline, or becomes, God forbid, you know, corrupt or gets censored by some government, which, you know, could happen. What happens in that case versus what happens in the lighter L2 case?

Vlad:
[5:49] Right, right. And I mean, I think, you know, we certainly, you know, we're, our team is, you know, we've always been friendly with, you know, Jeff and his team. So we certainly consider them good actors. But I guess in the scenario where something like that happened, they would probably do their best to compensate users. But it wouldn't be through smart contracts though, right? It would be something... Like kind of outside of that right so if that were to happen in or with any like other l1 right like if there was corrupt or hacked or just like like i don't know if they had an outage never recovered like you wouldn't be able to get your funds back through smart contracts like hopefully those projects and then we think hyperliquid is you know they're certainly good actors i mean they would try their best to compensate users in some other way but like but it wouldn't be done programmatically Whereas with our architecture,

Vlad:
[6:35] You can actually

Vlad:
[6:35] Like the escape hatch, like all those assets, you can't, no one can move money in or out between the L2 and Ethereum until the L2 can prove that what it's

Vlad:
[6:46] doing is correct and fair. And if it can't prove that, then within a certain amount of time, you know, users can just get their assets out programmatically through the escape hatch.

Ryan:
[6:56] Of course, this all presumes Ethereum stays up, but it has. It has a fantastic track record of that. It has a fantastic track record of security as well. Just the other piece here, verifiability. So I feel like we've talked a little bit about security, but there's a different concept here where you say verifiability of being in Ethereum L2 is important there. What does that mean?

Vlad:
[7:17] If I think about exchanges and What do customers need as far as guarantees, like kind of the hierarchy of needs there, right? I think like at the top, maybe at the top is like, you know, your money isn't stolen, right? Which certainly we've seen that, you know,

Vlad:
[7:32] FTX and other cases

Vlad:
[7:33] That can happen and certainly would not happen with, I think, you know, decentralized exchanges. Obviously, you know, module the point about the L1 or the L2 being secure, but certainly the design of the exchange makes sure that funds are safe. I mean, you know, self-custody, right? So that's the first thing. The second thing you want is that the trades you think happened are the trades that happened. So it's like kind of, you know, that there's, because that's not true for centralized exchanges, right? Like centralized exchanges, if you, like, if you think you bought and they exchanged things you sold, like, it's like kind of, you know, your word versus their word, you know, there's not really a way to prove which is correct, right? Now, that one, the reason I point that out is just to give credit to other projects like DYDXV3, like they had that, right? They proved that using kind of Starkware technology. What no one had done in the Ethereum ecosystem is the third hierarchy of needs, which is like when you send an order, it gets matched correctly. If you're an active trader, that's actually very important. Like it sounds like it's like academic, but it's actually very important because if that's not the case, especially during a high volatility environment, like the last few days, like if you're sending an order to buy it, 4200 and somebody else is buying at 4300 or somebody else sends in order to buy 4100, what if that 4100 actually gets matched before

Vlad:
[9:01] Somebody who's willing to pay 4,200, you know, you can do a lot of things. If the matching is off chain and not verifiable, like a lot of wealth can be transferred unfairly if that happens, right? So it's very important for the matching to respect like time price priority and kind of

Ryan:
[9:18] All the rules that our

Vlad:
[9:20] Circuits have. But also not only matching, liquidations have to be verifiable, right? Because again, liquidations, it's like, well, who's to say, you know, if the price falls from 4,200 to 4,000 and get liquidated well but maybe somebody else had the same exact position they didn't get liquidated like all those things also have to happen fairly so anyway both of those things are insured by the ZK circuits but again the L2 is only as good as the L1 it's on top of right like you can verify all that stuff but then Like, let's say those proofs just stay on our L2. Well, then, like, the whole thing can, like, as we talked about, like, the whole thing can just go rogue. Then, like, yeah, like, you can complain all you want about those proofs being wrong. But, like, there's not much you can do. Whereas here, the only way that assets can move back and forth between the lighter L2 and Ethereum is if Ethereum knows that the proofs are right. And that's kind of the key part here about the verifiability.

David:
[10:15] What you say, Vlad, about fair order matching, order execution, are these like theoretical concerns or are there examples in PerpDex history, in the PerpDex landscape of actually corrupted or unfair or untransparent order matching that has created just a bad environment for traders?

Vlad:
[10:35] Right. Well, certainly with MEV, that stuff happens all the time. I mean, I guess it depends on whether you look at the order book model or the AMM model. I mean, I think with Mavid, it actually happens kind of out in the open. It's not even a secret that stuff like that happens. I think with other order books... Yeah, it's a good question. I don't know if there have been any demonstrable... Because the thing is that you wouldn't know it, right? Like, if something like this happened, the only way you would know about it is if someone involved in that scheme, you know, if there was like a whistleblower or something. Like, if it is off-chain and there's no verifiability and there is... Well, I think we know that sometimes... Other projects in the past have had like special deals with market makers. We know that. But how far those deals went and how much did they hurt retail? I don't know if there's like evidence of that that's been out and open, but that doesn't mean it doesn't happen, right? Because if it... But it is a factor.

Ryan:
[11:28] It's a theoretical factor.

Vlad:
[11:30] Yeah, I mean, we know... Well, it probably does happen, right? It's just like, again, if you don't... If it's not provable, you can't really... Like, if it happens, there's no way you would even know about it, right? The only way someone would know that that happened in the past, if there was, if someone involved in that scheme kind of spoke out.

Vlad:
[11:48] Right, right.

Vlad:
[11:49] But we know that customers complain a lot about like why, you know, with other, with kind of, right, like even on centralized exchanges, like why did I get filled here, but this other trader didn't, you know, like anecdotally, you know, it kind of comes up a lot, but, and we know in Shrad5, for example, like the flash crash of 2010, like a lot of, Like there's a lot of really weird price action. And to this day, like I think Congress did an investigation like five years later and they still couldn't figure out what happened. So this stuff is like very important. And if you don't have it, then like the whole point of, again, this hierarchy of needs, it's like, you know, money not being lost. The trades you think happened are the trades that happened. And then everything that you did is verifiable. That's kind of why you want to use blockchain for this in the first place, right? But if you don't care about those things, then, you know, like you can just trade and TradFi.

David:
[12:41] There's a bunch more technical architecture conversations that we want to get into, like the ZK stuff, more about the capacity, the throughput of being a layer two. But we'd be remiss if we didn't also talk about the recent market volatility.

David:
[12:54] Just for context, for listeners who are less familiar with LIDR, LIDR launched a public mainnet October 1st. And Lighter at the Perpdex has been around for months. People have been trading on the beta version of Lighter for months, but the public mainnet happened on October 1st. 10 days into mainnet, one of the largest and most chaotic liquidation events across crypto occurred with $19 billion of liquidations happening inside just a few hours. Now, Vlad, when you build a perpdex, I think you know that this day comes eventually. I don't think you're ready for it to come 10 days into the public mainnet. So a pretty fast trial by fire there. Put us into your shoes during that time. When did it turn into an all hands on deck moment? Like how much of a crisis was it? Kind of walk us through what it was like to be Vlad.

Vlad:
[13:43] In that moment. You know, I guess like one thing to point out, and we wrote a postmortem about, You know, there was an outage that happened, right? But the outage actually happened five hours after the high volatility happened. So actually during the specific timeline you're talking about where markets got really crazy for about four to five hours, actually the system handled that correctly. There was no problem there. In fact, I think LLP also maintained liquidity. A lot of traders who would have been wiped out on other platforms didn't get wiped out. So actually the system worked fine during that time. Then after that kind of the database gave out and it's a little it's more of really like kind of like a bit of like the issue i think was kind of bad timing in the sense that we we we launched public you know we went from private beta to public mainnet like we upgraded some things to prepare for the higher you know kind of higher throughput new or more customers

Vlad:
[14:48] And part of that was kind of upgrading the database. We were planning to do that on Sunday morning on the 12th just as like a scheduled maintenance, right, and like upgrade the database. And we picked Sunday morning as the time that would bring the least disruption to the users, to the traders. And so the plan was to like do that database upgrade, take the system down for 10 to 15 minutes. But the fact that there was this huge volatility like Friday night obviously was not planned like had we known that we would have just done the upgrade like Thursday morning I mean they wouldn't have been happy

Vlad:
[15:23] Because the reality

Vlad:
[15:24] Is that users they don't like it if there's a two minute outage so it's like kind of like the perfect being the enemy of the good okay let's like time this upgrade exactly when there's the least activity and then A day and a half before that, we had the most active trading day of the last 10 years. And so, and again, it actually, database didn't die during that spike, but like five hours after that, it finally gave out. And we had to like, and upgrading it when it's unhealthy is different than upgrading it when it is healthy. So that's actually why it took like four hours to do that, four hours to fix the current thing. And then we still, we upgraded on Sunday as we planned to do. So now, now we're kind of the system has the infrastructure that it needs to have to kind of survive another event like this and with some buffer as well. But yeah, I mean, hopefully that gives you a little bit of the time because it's not, you know, it's like we, a couple of things happened earlier. Like us getting a lot more users happened a week before and then we were already planning to upgrade the database

Vlad:
[16:23] And then this

Vlad:
[16:24] Like volatility happened a few hours before it finally gave out. So I think I just want to make sure clear about the timeline because people I think have conflated some of the stuff.

Ryan:
[16:34] How long was that outage, Vlad?

Vlad:
[16:36] It was four hours.

Ryan:
[16:38] Four hours, okay. And this was somewhat after. So I'm curious

Vlad:
[16:41] Because it was like hour five through hour nine after the, it's like the high volatility was like the first five hours,

Vlad:
[16:46] Right? And then it

Vlad:
[16:47] Actually survived then, finally gave out. And so like, that's why we were going to compensate, you know, people affected during that four-hour blog. But those, you know, it'll be pretty manageable, right? Like if it had happened during the previous five hours, that would have been a much more complicated process to compensate.

Ryan:
[17:08] Right, right, right. Okay, so Leiter did have an outage, but it was five hours after. And you've talked about why. Right. So yeah, I feel like all eyes in crypto were on these perp dexes and perpetual exchanges, centralized or not. So sort of the three that got the limelight, one was finance, of course. There's a lot of mayhem going on there. Just finance the exchange and then also the perp side of things. The other was hyperliquid, of course, which did stay up the entire time, but had this auto deleveraging type of event that some traders kind of raised an eyebrow at. And the third, actually, given your volume and given the significance of size was lighter. And so I want you to give kind of a letter grade to these exchanges and like, or these perps, like how did they perform during the time when they're most supposed to perform, which is incredibly high volatility positions. So maybe start with your own letter grade for lighter, like what would you give yourself? And then we'll go to Binance and Hyperliquid.

Vlad:
[18:11] I'm probably not the right person to roll up those grades into a single score. You know the customers can do the you know it's up to kind of do their own homework on that but I can at least provide some input here I think whereas like if you look at the categories of reliability, liquidity and communication maybe if you look at break it up that way I would say speaking about us like clearly the reliability we did have an outage it didn't happen during the worst period but it happened so I would say in reliability it's like again just talking about October times, right? Not the entire, you know, I would say on the reliability part, it would something like between a C and a D, or let's just say D to be, you know, F would have been if the outage happened during the worst period. I will say like, D plus, C minus, something like that, right? I think on the liquidity part, we actually did great. I would say like, you know, B plus. The only thing that went wrong with the liquidity is that some of the market makers kind of left and it was really entirely up to LLP to provide liquidity. And so, you know, the market makers... And LLP is

Ryan:
[19:15] Sort of the insurance policy. That's the pool of last resort in the lighter system. Yeah, so LLP

Vlad:
[19:20] Is the public pool that is kind of the, you know, the market maker that's run by the protocol Anyone in the community can buy into it. And it takes, yeah, it's like, it's always there. It makes markets. It does process liquidations as well. If liquidations happen, it kind of takes over positions. And by the way, LLP does have ADL as well. It's just that the ADL policies are less aggressive. So it's not like ADLs didn't happen for any market at any time. It's just that we tried to provide as much liquidity as possible before getting to that point.

Ryan:
[19:56] And the ADL, for people who aren't familiar with these perps terms, that's automated deleveraging. That did not happen on your platform. That did happen on Hyperliquid.

Vlad:
[20:08] Well, I wouldn't say it didn't happen on our, it happened on our, like ADL isn't just something like, it's not like black or white. It's like, in some smaller markets and in some particular blocks of time, it did happen, but small, right? Like it affected maybe customers to the tune of, you know, tens of thousands of dollars versus like on some of those other platforms, the ADLs affect the customers to the tune of tens of millions of dollars.

Ryan:
[20:29] Can you define that more? Why is ADL in place in these purpose exchanges?

Vlad:
[20:33] Like why is it there at all?

Ryan:
[20:34] Yeah, why is it there? What does it do?

Vlad:
[20:36] And how did it kick in in a crisis? I mean, at some point, right? Like let's say that you and I are trading in market and like I'm long 50X and you're short 50X. And like, you know, let's say the market is trading at 100 and some catastrophic thing happens and the market goes all the way to zero. Like in like before anyone has a chance to do anything well like what would happen in tradfi right is like you would actually like you know like maybe you could potentially go after the person like claw back other assets but in defy you can't

Vlad:
[21:09] Like all i

Vlad:
[21:10] Have is the deposit i made in the wallet right you can't like get other funds from me or from you right so like like let's say that something like that happens well basically like kind of the total amount of like money in the system is now less than what would, right? Like European L now, let's say if that happened in this case, Like I put in $20, you put in $20, but now your P&L is $200. Well, like I only have that $20 to give. So like you're not going to make the $200 you think you're making. So now I'm not going to make money either, of course, but like I'll be wiped out. Like the whole system has to deleverage until like the money, the total amount of money in the system is like equal to the total amount that would be owed to all the different parts. And then the insurance fund can contribute to that, but it doesn't have... Like that, you know, the insurance fund is not big enough to like make everyone whole relative to what they're supposed to have. But then, but the point to avoid that though, you can kind of like do partial liquidations along the way, or, you know, you can just provide more liquidity so others can, can trade kind of in and out. So, so

Vlad:
[22:18] The ADL is

Vlad:
[22:18] Like the last resort thing that happens, but the reason it's needed is because of this example, right? Like, like ultimately, if the moves are strong enough and there's just not enough of a backstop, like you have to, because the total amount of money, you can't like call a bank or like write a letter to a lawyer and get other money into the system, like in DeFi, right? Like the amount that's been deposited to the exchange is the total amount of assets that's there. And that's not going to, you know, like you have to make do with that. And so that's why the ADL is necessary.

Vlad:
[22:45] Right.

David:
[22:46] Inherent to any leverage platform with margin, right? Then generally speaking, longs and shorts moderately balance out. And if price doesn't move that much in any direction, that's just okay. We're operating in the happy case.

Vlad:
[22:59] But if price

David:
[23:00] Moves violently and there's any discrepancy between a long and a short, then all of a sudden that gap can become more money than was ever deposited into the platform in the first place. And so you need to make sure that that doesn't leave some people hosed and others people not hosed. And so an ADL is like a equitable mechanism, equality mechanism.

Vlad:
[23:21] Yes. And all of these platforms have some version of it. I think, you know, like...

Vlad:
[23:29] I think some

Vlad:
[23:29] Of the discussions online have been, like, for people who've never seen it before, it's more like, wait, how does it, like, this thing doesn't... Yeah, they're learning

Ryan:
[23:37] About perps for the first time.

Vlad:
[23:39] Yeah, but, like, I think... Not us, though. Not me. Yeah, yeah, not you guys, right, but, like, I think some of the discussion... Like, because there's been, like, I think this turned into a lot of, there's been a lot of, like, bad blood online as a result. And, like, I think the differences between the different exchanges on this are, like, pretty nuanced, right? Like if you look at even our, you know, like if you go to our documentation and you read how the ADL works, it's like, you know, it's pretty, there's a lot of math in there, right? Like the differences between that and, you know, Hyperliquid or Binance ADL is like pretty nuanced. Now those nuances do matter and we can talk about that. But yeah, but the nuance, it's not like some have ADL and some don't. Like it's not, yeah.

Ryan:
[24:23] But let's talk, let's get back to the letter grade. So reliability, we got your letter grade for lighter. How about liquidity then?

Vlad:
[24:29] How'd you do it? Yeah, so liquidity, I would say, right, like kind of like B plus, right? B plus, okay. And the only reason I say not A is because like LLP, I think, did great with liquidity provision. Like I think the other market makers and I don't, you know, certainly don't want to throw them under the bus or anything, but like I think

Vlad:
[24:46] They kind of all left,

Vlad:
[24:47] Right? And I think they left not just us,

Vlad:
[24:48] But like other exchanges.

Vlad:
[24:49] I think like- They were very busy,

Ryan:
[24:51] Freaking out at the time. Yeah. How about comms? How'd you do on that?

Vlad:
[24:54] And comms, I would say we've done, you know, again, not to, you know, be self-serving. I mean, I would say like A minus, you know, I think we updated our users pretty quickly. And, you know, our customer support team was in touch right away. We had a form for folks affected, like we immediately announced compensation, that compensation will be given. We did a post-mortem on the technical outage. And now we're also going to do a write-up on what happened with LLP, which even though it did what it was supposed to do, like a lot of people didn't know that that's what it was supposed to do, right? So I think it's definitely worth a comms piece on that as well. But I would say, again, bias, but I would say like A- on the comms part.

Ryan:
[25:33] Okay, same for Hyperliquid. How'd they do reliability, liquidity, comms?

Vlad:
[25:37] Right. I mean, I think on reliability, I would say like, yeah, like maybe like B-plus in the sense that like you don't know outage, but there's definitely performance degradation and different I think certainly the high throughput and high volume was felt by the end user. But overall, pretty good, I would say. Yeah, something like a B+.

Ryan:
[25:58] Okay, liquidity.

Vlad:
[25:59] But on liquidity was the problem, right? A lot of people got liquidated who weren't expecting to. And I would say that one is maybe at a D there.

Ryan:
[26:12] That was because you think their ADL mechanism was too aggressive, you said?

Vlad:
[26:18] Yes. I mean, we're not experts in how their EDL works, but just from talking to customers who trade on both, that's, that's the, you know, that's, that's what we've heard. And also, if you just look at the numbers, right, like the 40 million that HLP made, like that 40 million didn't come out of thin air, right?

Vlad:
[26:34] Like that, a lot of, you know, people lost a lot of money. Now, some were just on the wrong side and should, would have lost money anywhere. But like, I think the, you know, again, I think others, I'm sure we'll parse the on-chain data and kind of figure this out. In more detail but just anecdotally from talking to folks like the liquidity really dried up quickly and they they were they had a hard time

Vlad:
[26:55] Getting in and out of positions and

Ryan:
[26:57] It's the traders really that are losing here it's kind

Vlad:
[27:00] Of that well the traders yes as opposed to who

Ryan:
[27:05] I guess i guess you know lps or something or yeah yeah

Vlad:
[27:09] I mean the right the lps are i mean it's a zero sum right It's like, I mean, ultimately, the traders have to, it has to be a good experience for the trader, right? I mean, all this other stuff, like having liquidity pools and having, you know... Like having points and airdrops and all this other, I mean, that's all good and well, right? But like, it has to be a good trading experience for the trader. Like if that's not there, then like all this other stuff is just smoke and mirrors.

David:
[27:35] There's a conversation going on that when I said Ryan and I are not, we're generally less informed about the perp tax landscape than the average trader, for example. So we're kind of like learning about these things in real time. There's a conversation going around on Twitter about the opinionated nature of the ADL system because every single per platform has to have an opinion on who gets demargined or deleveraged first. There's a queue of sorts. And I don't know to what degree that this is real other than this one tweet thread analysis, which seems pretty good. But it was talking about how hyperliquid LPs made money, where lighter LPs lost money, like lost a little bit of money. And what the interpretation of that means is that, well, lighter preferred traders over their LPs and Hyperliquid preferred their LPs over their traders, meaning because of the opinionated nature of the respective platforms, ADL system, ADL design, Leiter just allowed traders to retain more capital than their LPs and Hyperliquid was the inverse of that. To what degree do you think that that is an accurate statement? Is that reflected in what you just said, where you said you have to build a platform for traders first? Talk about the nature here. Yeah.

Vlad:
[28:51] So it's like, if you think about it, like, let's say you're running a casino, right? Like you're Caesars and, you know, people come and play blackjack. Maybe somebody goes on a great run. Now they're up a lot. Like, now does the casino have other, you know, maybe they're, you know, like, I don't know, maybe you have a credit line from, you know, BlackRock or something, right? Or Goldman Sachs, or like maybe, you know, they have shareholders and they have all these other constituencies too. And yeah, like you want to treat everyone fairly and

Vlad:
[29:18] You know, certainly

Vlad:
[29:18] The goal isn't to like screw over all these other constituencies,

Vlad:
[29:21] but at the end of the day, like the customer has to come first, right? Like, like if you're running a casino, somebody comes in and they hit blackjack and like, it's like the whole point of that is like, you get three to two when you hit blackjack and the casino is like, wait, wait, wait, actually, we're not going to give you three to two. We'll just give you, we'll just give you your money back. Like that, that's not, that really violates the kind of trust with the customer. Right. I think, so I think you kind of want to avoid that at all costs. It's better like it whereas if you know the casino has to like take out a loan at a higher rate for whatever you know with their creditors whatever that's like it's better to do that than have to then like violate the kind of the understanding with the customer so that's kind of how we think about right like the customer has to come first and then yeah you treat all the other constituents fairly as well like LLP holders like you know they had 15 up days in a row before the down day

Vlad:
[30:14] And LOP as a whole is up, you know, 220% year to date. So like, obviously we're not, you know, it's not like there was an intentional choice to, like for the stuff all works programmatically, going back to the point about verifiability, right? This stuff all, it's all encoded in the ZK circuits. It all works programmatically. And we made some design choices in the beginning that work this way that kind of favor the trader and those are all in the code. And so like, you know, it's not like we pushed a button and said you know let's like let's make this trade off in real time this all kind of worked the way it was supposed to work but the reason we feel good about the trade-off is yeah like you have to put the customer first let's

Ryan:
[30:54] Uh finish up that report card real quick so comms for hyperliquid

Vlad:
[30:58] I mean i think in general hyperliquid is very good at comms i think in this case i would say that uh

Vlad:
[31:04] You know, not the best though,

Vlad:
[31:07] Because I mean, I wrote something around like going back to the casino analogy, right? It's like, they were kind of saying, well, you know, we're, we're sad. People are going after us. But I mean, I think a lot of the questions that were raised were fair questions. And it's like, you know, they did very well, both in terms of fees and the HLP. So it's almost like, you know, there's this thing like no crying in the casino there. It's almost like no crying in the casino owner's box. Like, I don't know what there's to be sad about for them, right? Like they, but, you know, if they want to revise their leverage policies, they can then, or if they think what they have is fair, they can keep it going. Like, as long as it's transparent and everyone knows the rules of the game they're playing, you know, I think it's fine. So I would say comms there, like maybe like B minus, whereas usually their comms are in an A.

Ryan:
[31:53] Okay, okay. All right, how about Binance? What'd you give them on those scores?

Vlad:
[31:58] I don't follow Binance too closely, but just from all I know is just what I hear from, like, I don't trade on Binance anymore, right? I used to before, I guess, before probdexes were good enough to, you know, so I don't, I don't know firsthand, but Just from what we hear from our customers who are also on Binance, like I think on the liquidity side, it was also not great, which is surprising, right? But because you would think they have like tons of market makers and everything. But I think maybe like a C on liquidity side. And yeah, I think performance, some centralized exchanges had some outages as well. I actually don't know if Binance was one of them. I think, so I don't want to speak out of turn. I think some centralized exchanges had outages though. So Binance maybe just had degraded performance, you know, but I think there were some issues with like withdrawals, but I would say like not an A there, but I don't know, somewhere between a B and a C. I just don't know. But I think on the comms side, from what we've seen, again, secondhand, but like it seems like there was a lot of stuff about like the price of their token and like that's probably not the right time to talk about that.

David:
[33:08] Zooming out, Vlad, what have we learned? What have you learned and what maybe has the industry learned as a result of this cascading liquidation event? Like the microstructure that just builds around crypto.

Vlad:
[33:20] Anything novel to take away? Like the details matter, right? Like small differences in the math. Like I think most people when they see these ADL formulas, like, and I think we're guilty of this as well, right? In the sense of, yeah, like we stated what the formulas are and when we designed them, we favored the traders. But like it doesn't say that in the docs necessarily right like here you know the reason we made these trade-offs and not those trade-offs and you know it's like it was more just like here's the formula and we prove that that's what happens and that's It's correct, but like, I think a lot of the stuff

Vlad:
[33:57] Is like some of

Vlad:
[33:58] These really kind of technical, like super technical things, right? Like are kind of, you know, in, in, in the back of a documentation, you know, like a hundred page document of how the exchange works as opposed to really being like front and center and everyone needs to know. Cause I think that's like, this only comes up during like the crisis environment and then everyone forgets. And then the next time it's almost like the financial crisis, right? Like the credit crisis. It's like as soon as people recover, everyone forgot about what led to the crisis. So I guess the learning here is like details, like keep them in the forefront. Like I think whether, you know, we're certainly going to review our, the way that our mechanism works. Like, you know, we're already in the process of doing that, right? Because one of the complaints we had before was our liquidation fees are too high.

Vlad:
[34:47] Now no we don't now people see why that makes sense because yeah the fees are high where like maybe during like a big market move you pay a little bit more in liquidation fee than you might somewhere else but during a really big move the system LLP is there to protect you but again being more clear about that trade-off and like I think certainly I think there's a lot to be said about the verifiability piece too right like I'm like I think in a perfect world these systems should be audited they should be verifiable like i like teams like chaos labs for example that kind of like that's kind of what they do right like ideally it should be audited by like a third party like a chaos labs or you know other teams like like gauntlet right like like it shouldn't just be you know like the protocol design something and like publish some formulas in the back of 100 page doc and that's that you know so i think there are those learnings like that i'm still trying to process everything and you know maybe work with other folks in the space with for kind of more holistic solutions but I think the solutions it's about transparency it's about like verifiability but also about like decentralization in the sense that like like I don't think these decisions should just be made by like you know an engineer working on this thing who's like running a lot of math and they made a decision and that's that even even if it was the right decision like there should be more transparency, more collaboration.

Ryan:
[36:10] I mean, you've watched crypto markets for a while. This was, we've seen things like this before, maybe 2021, it was a kind of a big leverage sell-off then, but this maybe seems similar, but this was also an order of magnitude larger. Do you have any theories for like what caused this? Like, I mean, people are talking about like, oh, is this a coordinated attack?

Vlad:
[36:27] There's a few

Ryan:
[36:28] Pairs that they saw some liquidity on. Of course, there was, this was timed with the Trump kind of tariff scare. Is it just generally, you know, crypto has just so much leverage, this type of thing is bound to happen every so often. What's your theory of the case?

Vlad:
[36:41] Well, I guess before I say that, I want to come back, this is related, but I want to come back to your, like one of your first points, which is like, you know, this happened right after we launched public mainnet, which is true, but we have been running in private beta for eight months before. And actually there were some, like, like in February, there was a tariff thing too, right? And then like, I think like April 2nd, there was as well, there, there's some not, not as big as this but there's some pretty volatile moves uh and then there's some volatile moves on the upside right when like people like you know like tom lee and others started like buying eth like like there's like there's some pretty scary looking charts this year in general and for example like our system yeah like we had that outage five hours after this one but like all those other ones we we were able we handled without issue but but so i guess i don't want to over index on this particular crash like there have been kind of lots of them this year You know, I think what caused this one, I have, you know, we're, we spent the last 72 hours kind of just getting our own house in order. Like we haven't really been looking at like, you know, just looking at our own data with our own trading, making sure everything worked correctly, making sure customers who need to be made whole are going to be made whole. Like we've really been focusing on that.

Vlad:
[38:02] So root causes, can't say too much, but I've been in the markets for a long time. Like I think, I think the Napoleon quote does apply. like never mistake for malice that which can be explained by incompetence like I think a lot of times in the markets it is like that like it's like even like the flash crash of 2010 it was like I think the most kind of widely accepted explanationism or just kind of some like fat finger trades. And I think it could have been something like that here where, you know, was it a coordinated attack? I haven't seen any evidence of that. But again, we've been pretty busy just keeping our own house in order to do like a full, you know, a full diagnosis of that.

David:
[38:40] So this is a very interesting way to get mainnet up and running and just get a lot of stress under the system. But I kind of just want to zoom out and learn a little bit more about you and Leiter.

David:
[38:51] Where'd you come from? And also what made you decide to just build a Perpstex? Like what's your background and what was the motivation?

Vlad:
[38:58] If you zoom all the way out, like, and, you know, there's, first of all, like we have a really talented team. And so I really, you know, kind of, I will talk about my background, but I want to make sure it's not like I'm building all this stuff solo by any means. Right. So, but just my background, I guess I'm, you know, I'm not like a crypto guy that discovered finance. I'm more like a finance guy that went into tech, discovered AI, and then discovered crypto, and then built something in DeFi. It's kind of like the arc, right? So I started out, well, I did a lot of STEM stuff growing up. I was on the US team for the IOI and the IPHO, and did a bunch of math stuff too, and economics at Harvard. And then yeah so then i actually was at citadel and like when citadel first started looking at high frequency trading and then did that at another firm as well so kind of in total for eight years doing high frequency trading and trad fi through the so these crisis environments as i mentioned like crash of 2008 you know made off you know flash crash of 2010 i was there when all that stuff happened right and uh

Vlad:
[40:06] You know, then in 2012,

Vlad:
[40:09] I kind of shifted gears and came out to Silicon Valley and, you know, worked at a couple of startups, mostly in AI and fintech. That's also when I first read the Bitcoin paper, got really excited about that. Wasn't sure kind of what there was to do that was actionable, though, because from a trading perspective, it was very small. Like a friend and I built a bot in 2013 to try to trade Bitcoin. It was like, we calculated that even if our bot got every price move exactly right and traded at 10% of the market volume, we would only be making like $10,000 a month, like just enough to pay our, you know, like our expenses. It was like, you know, kind of put that aside, right? And this is when Bitcoin was trading it back.

Vlad:
[40:50] 30 or something, right?

Vlad:
[40:51] But anyway, so put aside, then another, later on, like another friend showed me more around like Ethereum and EVMs and that was really interesting. And then the thing that really got me excited was zero knowledge proofs though. And so 2017, 2018, like reading a lot of that literature thinking about what that means for scaling. So when my co-founder, Scott, and I started a company though, we considered a bunch of crypto ideas, but in the end went with an AI idea,

Vlad:
[41:20] Which we built.

Vlad:
[41:20] Called Lunch Club AI, which was like an AI that essentially like connects people. So it's like, hey, you know, Vlad, you're into DeFi and, you know, markets, you know, David, you're also into that stuff. You guys should talk, right? That's kind of how Lunch Club AI works. It actually still exists.

David:
[41:34] A dating app for friends.

Vlad:
[41:36] Yes, based on kind of professional interests overlap and kind of mutually, you know, then a lot of data went into it. So anyway, we built that and did really well during the pandemic. It's one of these startup ideas that like, because some ideas are just like obvious, it's not going to work and others work right away but this one like it worked pretty well but then it plateaued after the pandemic and so then we came back around and a bunch of our team me our engineers like a bunch of folks were like doing crypto stuff on the side anyway whether it's trading or reading zk research or you know doing you know going to ethereum hackathons this and that and like okay like let's really build something here and anyway going back to my lighter Well, we looked at, to me, like, okay, like, there are all these digital assets and I used to trade, right? And it's like, where would we trade them?

Vlad:
[42:22] Well, it's kind of weird.

Vlad:
[42:23] Like, why is 99% of the volume of digital assets traded on exchanges that don't actually use the blockchain to trade digital assets? So it's like, that doesn't really make a lot of sense because many, not all, but many of these assets, their whole point is to build blockchains for, you know, stuff like finance. But then when they're traded they don't actually use those rails, you just use centralized stuff which is less transparent than TradFi And so we kind of had that high-level idea. And then back then, when we looked at it, like the closest thing that existed to what we thought should exist was DYDXV3, as I think I mentioned earlier, and that they got a lot of things right, but the actual matching and the liquidations were still off-chain and not verified. So that's where, okay, we think we can solve that problem. And we worked on it for 18 months, launched the testnet last September and, you know, launched the private beta January 19th of this year. And then now we're in public mainnet. So why PerpDex though?

David:
[43:25] Because you could have built, I mean, Uniswap is a DEX. Uniswap does the job of kind of what you are talking about.

Vlad:
[43:31] Why do you

David:
[43:32] Care about PerpDex?

Vlad:
[43:33] It's not efficient, right? I mean, it's not efficient. Like I think that, you know, the AMM structure was needed at the time before Ethereum L2 was existed and before ZK was really ready for prime time. Like if you were if you just want to trade like directly on the ethereum on on on l1 like yeah the only way you could do it is is through an amm structure that's that's true but amms have other problems right like mev there is just slow right like if you're waiting 10 seconds to trade like that's not that that's not a good consumer experience so i mean and and candidly right that's why a lot of the volume was centralized right like it's not that all these people are like oh like I'm really excited about Bitcoin and Ethereum and Solana, but I want to trade all that stuff in a centralized way. Like that's not, it's more just because you couldn't trade it in an efficient way using its own tech yet. So like I think, because the cynical view is like, oh, people just want to gamble on this stuff. But I mean, you know, people can gamble on a lot of things. You can gamble on,

Vlad:
[44:35] Go to the casino,

Vlad:
[44:36] You can bet on horses, right? You can bet on sports, you can, whatever. I mean, people who trade digital assets, I think by and large, do believe in decentralization. So it's like, like, why aren't they using platforms that use decentralization? I think they were just not efficient enough to work for most forms of trading. Right. So, yeah. So anyway, so now why, why like perps specifically and not like, why didn't we build like a spot version? We actually did build one along the way. And the problem was that the ZK part of our stack wasn't ready yet for one thing. And secondly, we just saw like spot is really good if it lives side by side with perps. But spot on its own like you know like active you know if what you're building is the value prop is really great platform for active traders the active traders trade burps that's just the reality

Vlad:
[45:27] Like you know because

Vlad:
[45:29] Of it's just more efficient right like when you use leverage if you if you have a view on the market and you're a short-term trader it's just much more efficient to do so with leverage than not and so like spot markets also have a lot of uses more for just like doing swaps like kind of one-off things like oh i want to move some assets from this book to that book whatever right like but but if you're an active trader like there's there's not a lot of active traders would say hey like i'm a day trader and i trade in unit swap right but it's not really a thing like we've interviewed thousands of traders that that unicorn doesn't exist like it's like people who are active traders, they're trading perps. And so now, of course, once you have that, because you have to start with something that is... Like the thing customers want the most, and then you add other parts to it. So we're adding spot in a couple of weeks, in fact. And then we're going to add universal cross margin, and we're going to add the ZKVM sidecar. That's kind of the roadmap.

Ryan:
[46:22] I mean, it does seem like you're right. Like trading in general is a crypto's killer use case, but then perps specifically a highly efficient form of trading. So on L2B, it actually lighter is number six in terms of total value lock, the sixth largest L2 already, which is pretty impressive, you know, 10, to 15 days into mainnet. Also the number one, you know, layer two app chain, let's call it, because you're focused just on Perps itself. I do have a question here. So there's been a lot of discussion around, say, why Hyperliquid did an L1 rather than L2, or even why Tempo from Stripe or Arc from Circle or Plasma from, you know, kind of invested by Tether, why they're doing their own layer ones. And a lot of the answer comes down to, like a technical type of answer. I'm not sure if this is the real read between the lines answer or if this is correct, but a lot of them say, well, Ethereum is just like not performant enough. In particular, L2s are not performant enough. So I want to ask you the performance question. So it sounds like you spent a lot of time trying to, you know, make sure that Lighter had the performance necessary for high volume perps trading. So where are you on that? What is the, TPS throughput of Lighter today. What do you expect it to be? How scalable really is your architecture? Are there some limits to it and trade-offs?

Vlad:
[47:45] Well, let me talk about Lighter first and then I can talk about some of the other projects you mentioned and kind of the bigger picture of the industry, right? So with Lighter, our goal was always to have to build, you know, kind of the best decentralized trading experience, you know, in that, that you can have, which means low latency, low cost, verifiable matching, verifiable liquidations, and security, right? And it's like, you know, the security part we talked about, right? The escape hatch kind of being on top of Ethereum. The really nice thing about being Ethereum L2, and, you know, Vitalik talks about this all the time, right? It's like the more secure the L1, the more performant can the L2 be, right? Because when you're trading, like the L2 can be, first of all, the L2 doesn't need consensus. Since we use ZK proofs for verifiability, like anyone can generate the proof.

Vlad:
[48:41] And it doesn't matter if the proof was like, because the proof has to be correct and the rules for its validity are enshrined in the, like, you know, these ZK snark circuits that are public. And so like, you know, if the proof, like if our sequencer generates the proof and it's wrong, then that's not accepted by Ethereum or someone else can generate a proof too. So there's no need to wait. So basically what I'm saying is like, even if there was a decentralized sequencer, like whoever generates the proof first goes, like there's no need to wait for consensus. But I guess more to the point, though, the performance, essentially, like the L2 can be as performing as hardware can get. Like, there are no limits to the performance of the L2. The only limits are just economic, right? Like, because you have to pay, you know, someone has to pay to run those machines. And both the machines, too. There's, like...

Vlad:
[49:32] Know it's it's like you know think about like a race car that goes around the track and then there's like a big car that follows it to clean up but like and you kind of have to pay for the so the the race car is just actually doing the matching and like responding to requests within milliseconds and all that and then there's like another like bigger car that like generates proof that what the race car did is correct so you need to pay for the performance of both of those so now the latency part i think even right now later has the best latency in the market like Like if you try it out, like, especially if you're, you know, it is a centralized sequencer, but that sequencer is collocated close to where our customers are. And like, it's, it's pretty fast. It's like, you can feel, you know, it's sub hundred millisecond latency. Like it's, it's noticeable. TPS again is a matter of hardware and it's a matter of costs. Like, you know, I think the, like could later run at a million TPS right now. Yeah. Like we could, if we were okay, spending 10X on infra that we do right now. Now we don't want to do, because like right now there's not a customer demand for that. If there was, then the cost structure should be adjusted. Like we've thought about like if someone wants to do really high, like the kind of high frequency that has never happened on chain before, right? If someone wants to do like ultra high frequency, right? Like million TPS, then they should pay for that TPS directly. We haven't, that's not, you know, right now we have the free tier and the premium tier. There's no like pay for your own TPS ultra HFT tier yet, but that's something we're thinking about, right?

Ryan:
[51:01] So, what is the current kind of TPS? Well, the current TPS,

Vlad:
[51:06] Yeah, I mean, it's like around, you know, 5,000 or so is like normal. Like, during these crazy environments, it's spiked. I mean, I think For some blocks, it was like in the hundreds of thousands,

Vlad:
[51:17] You know, it's...

Ryan:
[51:18] Okay, so if you're getting to 5,000 already and you can scale to a million, I think for a lot of Bankless listeners, they're asking, okay, like, why are other layer twos different? Because while there's no sort of theoretical limit to the scalability of something like Base or Arbitrum, there's a practical limitation to it. And there, you know, Jesse from Base is working on it to get to giga gas levels of, you know, block-based throughput. Does that imply something different about your architecture versus some of the other more general purpose L2s?

Vlad:
[51:49] Our circuits are, like we did not build a ZKVM, right? We built custom circuits to run an exchange. And that's why later we'll also have

Vlad:
[51:59] a ZKVM sidecar for applications whose performance needs are a bit less. They can kind of be run in a smart contract that translates into circuits and then share a sequencer. to your point about the L2 beat, like we are an app chain right now. But I guess what you're getting at though,

Vlad:
[52:16] I mean, our design is,

Vlad:
[52:18] I think, very good optimization wise, but the other part is the economics, right? Like maybe some of those other L2s, you know, if they're used for stuff like payments, Like they don't have, you know, Citadel or Jump or Jane Street who want to use them for high frequency trading who would pay for that TPS, right? So like I think that's part of it too. So it's not just like,

Vlad:
[52:41] I mean, we

Vlad:
[52:42] Are very proud of our engineering, but I think also like this trading use case just has, there's just more economics to be put, to be brought to bear for it in the first place.

Ryan:
[52:50] Okay, so then this gets to the second piece of my question, which is like, well, how come all of these other platforms are opting for layer ones and citing performance reasons for that? Do you buy it? Or is there another motivator here? You know, those guys,

Vlad:
[53:05] They should talk to teams like Sysync Labs, right? And build a performant L2 on top of Ethereum,

Vlad:
[53:10] You know?

Vlad:
[53:11] I don't know exactly what made into it. Like, maybe, like, I don't know, like, maybe they made a technical choice two years ago that's not relevant today or something. Like, I'm guessing the stuff that's come out recently, it wasn't built yesterday, right? Like, I'm guessing they were building it over time. And a year or two ago, like, the building blocks on Ethereum weren't.

Vlad:
[53:31] As advanced and maybe they didn't want to,

Vlad:
[53:33] Like, I don't know if you're putting myself in their shoes, you start building in 2023 and all you have to go by is that, you know, Vitalik's vision, but the stuff hasn't been built out yet. And you're thinking, okay, it may take two years, but what if it takes five? Do I want to risk my business on that? Or do I just want to build an L1?

Ryan:
[53:49] I don't know.

Vlad:
[53:50] That's probably the thinking that went into it, but I don't know.

Ryan:
[53:53] Another explanation I've had, which is just like, it's not a technical explanation, but I'm wondering, so LIDR does not have a token right now. Maybe it will at some point in the future. There does seem to be something that I've called before the L1 premium, which is for whatever reason, if your token is an L1 token versus an L2, in these current market conditions, there's some sort of premium associated with that L1 token. Maybe it's just the simple logic of, well, if it's an L1, then we can take Ethereum's market cap, which is like $500 billion. dollars, we could be a small sliver of that market cap versus taking an L2 market cap. So it's possible that the explanation could be it's higher ROI. You have the option and you have the L1 premium if you do an L1 rather than an L2. And that could be some of the explanatory reason why. I don't know if that's COPE or if like there's something I'm missing here.

Vlad:
[54:45] That's not really based, like I'm a first principles thinker, right? As I think you guys are too. Like to me, that's not really based on any like that narrative probably is a thing, but it may be, I don't know, like maybe some famous VC said at one time and everyone was like, oh yeah. Like, it's like, I don't think that's based on anything, right? If you think about how other things work, like who, like, you know, I don't know if Ethereum is like HTTP, like, or whatever, like, you know, or like Cisco, you know, there's a value in their other applications on top of that, like Google and Facebook, like they recruit a lot of value too. It's not like, you know, Facebook built their own version of HTTP. You know, I just don't think that really, like I think you have to look at each application from first principles, like who are the customers? Who, you know, what are the risks? Like where does value accrue long-term and stuff like that? I'm not just like a VC meme.

Ryan:
[55:38] From a first principles perspective, so you've talked about all the benefits of being in layer two, verifiability, security. You have to maintain a consensus layer. You don't have to inflate your token in order to pay for validators, all of these things I imagine factor into it. Are there any downsides, though? Do you lose any sovereignty being an L2? Is it maybe harder from a technical perspective? Are there liquid?

Vlad:
[56:00] Yeah, I certainly don't want to belittle kind of the technical efforts of other teams. So, you know, but like maybe some of the projects you mentioned, the reason they didn't follow the kind of path we did is like, they just didn't have the in-house expertise to do it. And that's why I mentioned teams like Sysync Labs and others like Axiom that are helping developers kind of build efficient L2s on Ethereum using ZK. But the sovereignty thing, yeah, I mean, I think that's probably, I mean, of all the arguments I've heard, there's probably something to that, right? Like being reliant on Ethereum. I mean, I think if one of the risks is like, well, what if Ethereum itself becomes corrupt? Like that is a risk. And if that's the risk you're solving for, you know, that could be a reason to make a different decision. Right. I mean, I think that, that is, that is real. I mean, but I think, yeah, like keeping the technical stack easier, more under your control,

Vlad:
[56:57] Reason too, right?

Vlad:
[56:58] Like, we're definitely, like, are we somewhat at the mercy of the, of like, like, if Ethereum, stuff happening on Ethereum was like two years behind, would we be where we are today? Like, probably not, right? So, so there are things like that, but, but, you know, we're very aligned with the Ethereum community. And, you know, like, I think that that's certainly a bet we've taken and we're, we're, we feel good about it.

Ryan:
[57:20] When you say harder, it's interesting, right? Because there's, like, it's harder to do an L2 from an engineering perspective, there might be lots of reasons why. It also seems very hard to like bootstrap your own validator set and work on like decentralizing that and making that kind of the consensus layer. Like that's a whole nother moving part that L2s don't actually have to contend with.

Vlad:
[57:39] I mean, is it though? Like, I mean, I think there are frameworks to do with like Cosmos. Like, it's not like they're not also like...

Ryan:
[57:45] That's been done many more times, you're saying?

Vlad:
[57:47] I mean, it's not like no one's, you know, like what we've done has never been done before. I guess the closest was DYDXV3, but even they did it in partnership with Stark we're not alone hmm

Ryan:
[57:58] I get it.

Vlad:
[57:59] I mean, none of this stuff is easy, right, by any means. But in terms of the degrees of how hard it is, I think what we've done is kind of right at the top. Maybe we can place

David:
[58:08] Ourselves in Leiter's roadmap. So Mainnet launched 14 days ago, biggest liquidation event ever in crypto, 10 days later. But that wasn't part of your roadmap. That was an external force. What happens next? What are you guys working on next in terms of your roadmap? What's the grand plan and.

Vlad:
[58:26] Where are we? Right, right. Right. Yeah, I think I quickly talked through this earlier answer, but let me expand on that a bit. Right. So, you know, Spot is we were planning Spot by kind of end of this month. You know, the recent events might have set us back two or three days, but still like end of October, maybe first week of November, still scheduled for Spot. After that, you know, if you think about kind of core infrastructure, then there's like things like the ZKVM sidecar. There is the idea of a universal cross margin, right? Where any asset.

Ryan:
[59:03] Just real quick, Vlad, that the sidecar. Are you referring to basically besides just being a perps app chain, having more general purpose block space and more of kind of a platform type play?

Vlad:
[59:15] Yes. Yeah, yeah, yeah. So you could write, because like right now, let's say someone like an Aave wanted to like, first of all, we already can integrate with the stuff that happens on the L1 on Ethereum already. But let's say someone like an AVI wanted to like, or whoever wants to write an application

Vlad:
[59:32] That runs directly

Vlad:
[59:33] On the L2, like they would have to be writing ZK circuits right now, right, or, you know, that's, but with stuff with a ZKVM sidecar, they can actually just take their existing smart contracts, they can run in the sidecar, but that shares the sequencer with the perps exchange. So then you can kind of like,

Vlad:
[59:49] You know, like those

Vlad:
[59:50] Assets would can be used as collateral or used for liquidations so on so that's the zkvm sidecar thing and then there's the universal cross margin which is pretty cool right like imagine that like all these people who hold eth like imagine you can use your eth as collateral when you trade and of course like think about what just happened like liquidations like you'd have to be pretty conservative like maybe the leverage you take on if you do that isn't 50x maybe it's 10x you know but like but it's still better to have some version of that than not especially if you know instead of having to like force people to use usdc as collateral and then

Vlad:
[1:00:27] Yeah, beyond that, just on the markets.

Vlad:
[1:00:28] I write different types of markets, pre-launch markets, real-world assets, options. All of that is stuff for the next few quarters.

David:
[1:00:38] So you're looking to become kind of just a financial super app, super platform, where any financial instrument can be turned into a ZK circuit and added to the platform? That's correct.

Vlad:
[1:00:50] Yeah.

David:
[1:00:51] Wow. Has it always been that ambitious?

Vlad:
[1:00:54] Yeah. I mean, it's always been, like I said, markets, you know, when we started, it's like, why are all these digital assets traded in ways that don't actually use the underlying tech? And let's build, let's build, you know, a way to trade digital assets that fully leverages kind of decentralization, verifiability, security. And well, and the scope of what digital assets means has grown a lot while we were building also, right? Like, for example, real world assets weren't really on a roadmap in the beginning, but now with kind of new, you know, new administration, new policies, now they are, right? Or things like, you know, like prediction markets maybe as well, or things like options. So all of that, a lot of that is driven more by customer demand though, right? Like we're not, we're not just like building stuff in the lab and like seeing what sticks. Like we actually, we're pretty transparent about it. Like we ask polls, like, okay, what should we build next? And the one that came up was like pre-launch markets. So we built pre-launch markets.

David:
[1:01:53] So the perp-dext landscape has this huge... Hyperliquid colored glasses around the whole concept, largely because of just the absolute success of the launch of the hyperliquid token. No investors, no VCs, a very generous share given to the community, and then the token just ripped afterwards, did a 10X. And now I don't think you can be like a perplex in the space and not really have that fact, that series of events, kind of just in the back of your head. Lighter has a points program, Presumably points turns into a later, lighter token.

Vlad:
[1:02:30] Any details?

David:
[1:02:31] How do I get points? When will the token drop? Any sort of details about this aspect of a lighter?

Vlad:
[1:02:39] That's an easy one because we're going to be publishing. So as season two starts, we're going to be publishing the formula for how points are generated this week.

David:
[1:02:49] Why has it been a secret so far?

Vlad:
[1:02:51] Well, so season one, the formula used in season one was a secret. Formula for season two will be much more transparent why the secret because you know like we didn't have systems in place to at that point to prevent bad actors from gaming the system right so also i think there's something interesting when it's like new people experiment like if they don't know what it is and they try different forms of trading you kind of get like a better stress test for your product people

David:
[1:03:19] Are poking around in the nooks and crannies.

Vlad:
[1:03:20] Of the protocol right Right, exactly, exactly. But now we know, you know, we have like 250,000 traders, right? So we know what kind of stuff they're going to do. Like, it's unlikely that a trader will come at this point and discover a whole new way to use, you know, to use Lighter. So, like, I think we can be a lot more transparent now. And the transparent, like, why, like, the argument for the transparency is that especially, like, institutions, you know, market makers, HFT shops, like, they have quant teams, like, run strategies. Like for them it's very helpful when they can actually say okay like when I run the strategy and then like you know this is how many points I'm going to get for the strategy and like

Vlad:
[1:04:00] That adjusts their models,

Vlad:
[1:04:01] Right? And that's good because they can, for example, provide more liquidity if they know they're going to get points.

Vlad:
[1:04:07] Again, like maybe if they had known that they would have stayed in more on the 10th and we would have had even more liquidity than just LLP. But your question about like, yeah, we're thinking about kind of year end as kind of to wrap up season two. And, you know, I think you can expect kind of something to happen around that time, but we'll kind of make announcements as we get closer. And I think, yeah, I think your other point was about like, you know, the market... You know, market comps to other projects, right? Yeah, I mean, again, it's like, to me, this is all about blocking and tackling. Like, you have to treat the customer right. You have to have reasonable policies, right? To be transparent about communication. Like, this whole thing about like VCs, well, we have some very helpful partners on the capital side. But for example, you know, Founders Fund is their, like their whole ethos is that they are there to help when you need it. Otherwise they get out of the way. And so it's like the founder kind of has, You know, the founding team, you know, makes kind of the core decisions, right? And for us, we, of course, get input from our customers when we make those core decisions. So it really depends, like, I don't know, maybe people have been burned with some particular VC-backed projects before. So this whole narrative appears, but it's like, from our perspective, it's no different than,

Vlad:
[1:05:24] You know, if someone on a team won the lottery, you know, won the Powerball and had a billion dollars to inject into the project, it would be no different than the VC partners we have. Actually, it's even better because they can help when we need them.

Ryan:
[1:05:38] Vlad, you guys premiered on L2B, which is sort of a stakeholder that looks at the property rights assurances of all the L2s out there. And you got four out of five pie slots green. It's pretty impressive to be still a stage zero. What's the plan to get to stage one, stage two? Is that a priority for you?

Vlad:
[1:05:58] Yeah, yeah. I mean, that's kind of for the team that's working on the core infrastructure, the cryptography side. Like, yeah, L2Beat is, you know, one of their core metrics, one of their core KPIs.

Vlad:
[1:06:12] So I think we should be all green pretty soon, but I guess you'll see it when it happens.

Ryan:
[1:06:18] Very good. Vlad, thank you so much for joining us today. This has been fantastic. We appreciate it.

Vlad:
[1:06:24] Yeah, appreciate the questions.

Ryan:
[1:06:25] And for everybody out there, none of this has been financial advice, of course. Perps are risky, so is the rest of crypto. You could lose what you put in. But we are headed west. This is the frontier. It's not for everyone, but we're glad you're with us on the Bankless journey. Thanks a lot.