Sei: The L1 Built For Trading | Jay Jog

Transcript

so at a high level we're building the best infrastructure for exchanges the question then becomes what does that infrastructure look like what's up everyone welcome back to another episode of Xerox research today is March 27th and we have a great interview lined up with Jay the co-founder of say which is a high performance Cosmos chain designed with dexes in mind but before we get into the interview as always we are joined by two blockworks research analysts Xerox pibbles and ren to discuss the latest Market happenings Ren I'll kick it over to you first for your hot Cedar cool Throne sure so I have a hot seat this week um in the haasi lands on Gary genser and SEC so five days ago SEC Sunday was noticed to coinbase well it's noticed is

basically a request for information for an ongoing sort of like ego process and it basically is the SEC going hint you're in trouble you can kind of own up now and provide us with the information that you think you're in trouble for or you know including basis case they just say screw you we're gonna ignore this and we're gonna fight in court so a coinbase identified the wells notice was regarding an undefined portion of listed digital assets the coinbase earned staking service so not coinbase if that's something else coinbase Prime and coinbase wallet but in reality the world's notice gave coinbase zero Clarity on basically what type of infringements they had right

um and after that we had a bit of discussion on Twitter coinbase's Chief legal officer Paul will stated that coinbase met with the SEC more than 30 times over the past years they described their staking and asset listing process is free the IPO which the SEC approved and they've gone a few regulatory compliant vote they've acquired two broker Duty licenses in 2018 obtained a DCM and dco license from the cfcc and even shut down coinbase land towards the end of 2021 so this action from the SEC is pretty disappointing because they've provided zero feedback they've given zero regulatory Clarity and they've developed zero regulatory Frameworks for crypto or digital assets and so they're

going full enforcement and cracking down on coinbase even though they've been one of the most quote-unquote well-behaved players in centralized exchanges within the space and furthermore they've cracked out pretty hard on a lot of like stable coins for example pack stops they cracked down pretty hard on crack in the staking program and it really makes you wonder what the end goal is here right do they just want to shut down everything crypto related um within the United States this guy against or just a Bitcoin Maxi and doesn't want anything centralized but yeah the SEC is definitely in the hot seat right now yeah I guess the ultimate question becomes like this to me I don't know to me this feels like the the last battle right like if coinbase goes down

it does start to look pretty Grim at what's left and like how we build from there but um I really liked the way that the company and specifically Brian Armstrong is kind of positioning themselves to to really take on this fight like hey we don't believe we've done everything anything wrong like we made an effort to comply with you we were in a publicly traded company on an exchange that you oversee you've you've read through our financials you know what we do like we want to comply you're just making it unreasonably hard for us to comply um and like to be if you're just like following sound reasoning like that seems like a good argument and like it makes you I feel like that makes the average person want to defend coinbase

to be honest it's like it just feels like the classic story of government overreach and like there they seem like the good guys trying to do the right thing um which I guess like is super contrasting like the finance news you hear today where it's just like yet another example of for some reason it's so hard for these exchange CEOs to like not trade against their customers but I'm sure we'll get to that later but I I just I bring that up just to say like you know coinbase really has tried to do everything by the book and so it seems especially relative to its peers and this again I I just I don't know it feels like the last battle and Brian Armstrong had a good tweet Thread about how they're preparing to fight this and you know they believe what they've done is right and they've been by the book

and I was like I had a I had a response to it I was like all right like I'm ready to suit up where do you need me like General Armstrong you're my guy yeah I don't have a whole lot to add to it it's it's just an unfortunate situation and we just gotta sit sit tight rude on Brian and uh hope that his baldness prevails um but yeah I think just in in general speaking talking it's going to end up being a good thing once we do get the regulatory Clarity that's when the big money institutional Capital will actually be able to play around in the space we'll actually know what's right what's wrong what's taxed this way what's text that way uh and and I mean it's it's regular regulatory Clarity that we've been seeking for the last two three years at minimum so I'm just

excited it's actually starting to get underway and I'm just going to hope for the best outcome yeah I know I think that makes a ton of sense um and again like if it feels like if we can climb this mountain then getting that framework kind of sets the precedent of like here's how you comply in the United States and assuming I I guess assuming the coinbase win it kind of pushes Us in that direction um but yeah I'd love to hear somebody talk deeper on the binance side of things does anyone have them in the hot seat this week yeah that's actually my hot seat and it's kind of ironic because they're getting or sorry cz's getting sued by the cftc which is the uh us domicile agency that oversees Futures and commodities trading um so basically they're suing CZ for

offering investment products like options Futures to us customers which apparently isn't isn't allowed in this jurisdiction um and there is a lot of different messages on Signal and encrypted messaging app between the company and apparently their Auto delete conversations and that they were also trading on personal accounts on the exchange that they you know run which is clearly a conflict of interest that does not sound good and this comes right after you know all the other regulatory Crackdown we've seen in the crypto industry and paxos with busd um you know there's some rumors that maybe this might be tied to USD T in some way I think that's like the absolute worst case scenario and I don't think it's plausible or likely I should say but I think the most notable thing

to take away from it is the fact that the cftc is trying to regulate crypto at the same time that the SEC is who is you know Securities and the cfdc specifically called out Bitcoin eth Litecoin and a couple other stable coins calling them Commodities so hopefully the cfdc gets their way and and wins this territory and gets to regulate it honestly because it seems like they're a little bit more friendly than the the SEC and Gary Gensler but that's kind of the TL tldr on this situation yeah I guess like the first thing that pops into my mind on that is like where is there any contagion from this you know I feel like we're also scarred of the last 12 months of like one bad thing happening leading into four other bad things

um so like first thing I did was pull up Dune and head over to Tom Juan's uh Finance flows dashboard and they've experienced around a 500 million dollar outflow today uh which sounds like a lot on the surface but if you flash back to the FTX drama in November 22 there was a single day I believe it was the 5th of November uh 2022 and there's a two billion dollar outflow in one day so on on the grand scheme of things like this is kind of another day for binance which is good right because if there was some malpractice going on over there and there was all of a sudden like large outflows for consecutive days in a row then I would begin to get concerned of like okay well if they kind of were playing around in you know a couple the

bad pools of water like were they not backing assets one to one is this going to lead to another FTX like situation but the early signs is like not a mass rush to the exit which is at least a positive development in my mind um and then I guess the other question is like if we lose binance what's the impact on the industry and I guess I'd kind of pose that question to you guys like hey do you think binance goes down and B if so like how does that impact the growth of the industry yeah so I have a couple thoughts on this first I don't think binance is going down off of this even though CZ tweeted for today which is associated with death in China I believe um but I think that they're going to go

after binance for having Futures Trading open to U.S customers in like 2019 but I think that they they really circumvented that in 2020 2021 I mean we can't log on to the binance website right now and go trade like 50x Leverage so like they have complied it just took them a little while so I think if anything you know maybe they pay some fines but uh in two months I think uh we'll be over this yeah I would end on that I would imagine Financial CC has a pretty big award chest that is a lot bigger than what we know of

um and even if they settle or they have to shut down like everything that they did wrong or even if they shut down the apparent 300 like internal accounts that he used to trade under exchange against their customers like even if they shut all of those down at the end of the day Finance is still a dominant Central exchange they have probably more than 85 market share and they are still a very profitable business as a whole right even if you take out the US portion of your business they make a lot from Trading fees from other countries all over the world and probably one of the only other like companies or products that has found product Market fit right what is finance institutional Finance retail uh digits on binance Smart chain like there is still a lot of user

activity there that won't go away anytime soon I believe yeah agreed Ren and I guess just in terms of Market structure though they are where they are the place where a lot of people in the world go to trade so if something would happen to binance like it would obviously really hurt the amount of liquidity available for trading and then I would expect more volatility um just in the space in general so whether that's to the downside or the upside just less capital is going to move markets in a lot more drastic of a way so that's something to pay attention to there are two other interesting things that I got out of this one was that they the cftc got a hold of like these uh telegram or signal messages that were and like knew they were set to Auto delete so

my question is like how uh if to me it feels like either there was like an internal leak that happened and somebody kind of like brought this the CFC from the inside which would be kind of interesting uh or B like did they somehow Force signal to give them those messages um so yeah I don't know like does that does anyone have any takes on how that would happen and then the other thing is I guess just the size of the fine is then like if they do settle out and don't take this to court um it'll I would assume would be like a massive fine you know that like kind of like leads to CZ not you know having to admit guilt in any way and then Finance keeps operating uh so the question is like is the size of the fine something binance can pay and one of the uh points

in that was like any Us customer affected by a liquidation they're like they can be fully reimbursed for that loss because it never should have happened which is pretty pretty interesting um so if that's true I mean you know we were all trading over the last year and throwing a 50x leverage uh trade on in a down only environment it pretty much ends one way so if they have to repay every liquidation that happened uh to customers that were based in the US like that could be a pretty significant sum of money yeah I would say that um towards the first point in terms of like signal signals in a VA for the encrypted peer-to-peer app and it's probably the messaging app you want to use if you

want to Dean on like the Privacy and secure side of things I wouldn't be too surprised if it was a disgruntled internal employee because even though Finance is like a four thousand five thousand dollar organization I have heard that work culture is pretty crazy and it's pretty intense like super long days and so yeah I wouldn't be surprised if it's an avoid essays okay I've had it with CZ I've had it with this company I'm gonna link it because I'm not sure how else the SEC would get Auto deleted messages now and I would assume they were pretty like conservative with their division timeline maybe it's like all of these after 30 or 60 days um so yeah as far as the second Point goes I don't think the final will be substantial enough that it will off

mine ends and even in the hypothetical worst case where the fight is large enough to completely deplete their cash reserves right I think they'll just continue from there continue operating as a business because they are so profitable right now even in the bear Market I would guess um so I think that's fine for them yeah definitely I can go next I have a cool throne for Oiler Euler Finance I'm gonna call it Euler I know after Capital let's call it Euler but I think that sounds weird um yeah so they got hacked for 200 million uh it was a week or two ago and um that was a pretty devastating hack for a lot of people

um what's interesting is that whoever hacked it at sent an on-chain message and they're like we're not going to keep funds that aren't ours and also he did some some weird stuff like I know he sent some fines to the Ronin Bridge exploiter who um who we thought had something to do with like the Lazarus Lazarus group the North Korean hackers but um I don't know he could have just been trying to uh muddy up the trail but what's really neat is over the weekend they got like 5800 East back and as of today on Monday they got like another 30 mil back so they've recovered like 130 million dollars so far so that's just things you love to see

because uh the Euler team's a bunch of great guys and it's I think it's pretty critical for defy that um we figure out how to get these funds back to the affected parties yeah the thing that was weird to me is just how long it took them to actually send the funds back and then like also like you said he was like he sent like 100 each like three or four different wallets of different prominent like tags so I'm not really sure what the logic there was I'm sure he just he she they came to the realization that I can't do anything with this money I might as well just return at this point and that's why I don't really understand these hacks in the first place like I guess like teaching people a lesson but like

why not do that and just submit a bug report and hopefully get a nice Bounty I know some of those bounties aren't always the best but it's better than getting thrown on someone's radar in terms of like getting in in a lot of trouble and arrested in your your own jurisdiction so yeah surprised to see it but also happy to see that the funds get returned I agree with you Spencer that uh Euler is a good protocol and they were just trying to build a cool new Innovative product and it just kind of ended unfortunately yeah and the way it happened as well was just super unfortunate I was like they were very rigorous in their audit structure and there's like one contract with one piece of it that wasn't audited and of course that was like where this bug was living so really just an unfortunate break for

them and like the question is how do they bounce back like how do you guys view protocols that have had significant hacks in the past but are like you know have since patched that bug in our continuing forward is that a protocol you would use again or does it leave like a black a black mark on them hey they can't get hacked again right and Rory Finance get hacked like two or three times what's the other author chain Thor chain got hacked multiple times yeah door chain had some bad beats as well for sure but like they are I guess the counter argument is you know if you do squash out some of these major bugs then like not to say how many can be left but uh you know that I don't know like it's like I could kind

of see you as that being like a hardening process right like the longer you're alive the further you're battle testing and that like really proves to be true but it's like I guess you kind of start over at Ground Zero now and you lose all that previous like battle tested uh Acumen that you've built up I want to say that it's I lean towards it being more footage actually obviously it sucks that right now that's just 70 million dollars that are lost out there but at the end of the day if you do have a good product there will be new participants that come and save two or three years down the road and chances are they will have no clue that you'd ever got hacked before right and they're gonna think wow this is like a great new lending volume protocol it's really cool

I have different like collateral categories and I'm gonna use it so at the end of the day it's so how good of a product you are obviously that's a very simplified statement um but I lean towards it being a good event for you and I think they can bounce back yeah it's a good reminder too just like you can't just look at tvl and like time in the market and think all right I'm completely safe I can eat my entire balance of stable coins into here to earn three percent four percent whatever it was offering like there's always risk no matter where you put your money in crypto so it's it's kind of a nice wake-up call and reminder on that front yep sure there and I can take us home here with uh with another cool throne and this one kind of ties into

what we talked about last week arbitrim claimers the airdrop is still pretty white hot um right now token price is kind of settling around a buck 20. uh making the total airdrop value about a one and a half billion dollars that went straight to users of the of the chain so yeah again in total is about 625 000 eligible addresses which there's been some interesting uh data done on like what percentage of those were cyber attackers and seems like the general estimation is about 20 of those 625 000 wallets uh seem to be like related to cyborg entities that I kind of like slipped through that work done by the arbitrum and nanson teams to kind of prevent that which I feel like that's

kind of like an impossible bug to solve and like I'm no civil expert by any means but it seems like you know everybody was CyberLink this for the exact same reason like you're giving out a billion and a half dollars of free money I'm gonna see try to see how much of that I can grab was kind of like the logic so you know they they did deter a good number of those but it does seem like there is still some analysis showing that there there were still some cyber attackers that I received air drops um so nonetheless the you know token price has kind of based out around a buck 20. I was kind of thinking fair value in the range of 80 cents to a buck 10. so it still seems to be in like that fair value range it puts it you know

slightly above optimism in terms of in terms of market value and you know based on the tvl and daily active users that seems ballpark Fair um you know it's got a much much more vibrant native defy ecosystem and you see that like still really having a having an impact right so if you think about an airdrop like you want to reward users for using your chain but you also want them to continue participating in the participation of the chain and net inflows are still in for stablecoin inflows meaning users are still bridging value to uh the arbitrim chain and you also see things like on-chain volume still very very high so you know generally the ethereum is always really the The Hub of Unchained uh Dex volume

and you usually see like BSC coming into that number two slot and right now you see arbitrine by a mile so over the last seven days there's been about 12 billion dollars of Dex volume on ethereum and 5 billion on arbitrim so you really see that like change of pace uh and people really are trading arbitrum the token on the Chain as well as like really promoting like just broader defy uh participation in the ecosystem so it seems like it was a fairly successful airdrop in terms of in terms of rewarding users and getting eyeballs on the Chain um and you know it'll really be interesting to kind of like take that next step of okay well we've made this token how do we like you know put value towards a token give it utility uh and

like really that logical progression there is through governance so they did launch the arbitrim dial and when you claimed your airdrop there was sort of like this Auto step to delegate your token um and so you know Matt I would you know as we move forward we're going to see a little bit more involvement on the governance side of things and that's like a gradual step towards decentralization so exciting to see the focus there uh and then of course the The Humble show here uh if you are you are an arbitrim token holder and are interested in delegating your tokens be sure to delegate to blockworks research uh you know we really believe in the promoting growth of this ecosystem uh and excited for the opportunity to kind of hold that governance responsibility so uh if you go to our arbitrim dashboard at the very bottom there's an

arbs delegation section and feel free to check out that and there's a couple links you can you can click there yeah I would take the other side of this one myself I think the airdrop was a really poorly designed airdrop like I had three or four wallets that got a drop and I hardly use like two or three of them I had friends who did one thing on arbitrum and got an airdrop as well immediately sold it of course in order to just cash out and put it in the bank so um yeah I don't know think they should have done something I don't know what they should have done but maybe just like looking at people who got liquidated on GMX or something or like different things like that like I think the the methodology was a little too straightforward the only one I thought

had a lot of credence was uh Nova Bridgers which is their kind of like data availability committee chain for gaming and social applications so I I didn't actually get that one too so I'm I'm actually unbiased here but I didn't think the actual allocation was super effective I'm curious if you guys agree or disagree I feel like you could also make the counter argument to that because there wasn't a whole lot going on on Nova so like the only people who went there were people trying to farm the airdrop because like oh let me check this box so I think that really just proves like how hard it really is to to kind of weed out like who's deserving of this token just uh like it's obviously very easy on a case-by-case level but you can't do that with you know around

625 000 recipients I think what would have been nice and very appreciated is an extra reward for the people who bridge to arbitrum day one and we're farming Arbor neon I got rugged on carbon like those were the good OG days and um you know it breaks my heart to see someone who plays with dope X and GMX getting air dropped more than me they weren't there for the day once obinian was definitely good times I think another thing about the airdrop that robbed me the wrong way was that they included previous civil attackers in the list of civil attackers for this airdrop with me said for example if you save out

another airdrop such as hot protocol they use their list and include it in an included that list for this air job and personally I feel like sure if you like civil attacks something in the past you shouldn't be eligible for that previous air job but this is like a separate instance you know like if you didn't sample the object from airdrop you shouldn't be penalized just because you save out and previous air job but that's definitely like a very personal and controversial take and I would like to hear the different opinions oh wait Ryan did you say that they did like exclude people from the airdrop who had previously cycled other protocols airdrops yeah they did they excluded previous civil attackers for other

airdrops okay see that makes a ton of sense to me like it's just like the same as like a a loan if you go to like a bank or something like they're gonna look at your previous credit history like if you've done well paid your bills Etc then you're going to get a line of credit but if not then you're not so makes sense to me but I could see the other argument as well yeah I'm in your corner here Sam it feels like I don't think this is probably not the first time but one of the larger occurrences of like your on-chain actions are creating an on-chain reputation um and that's like you know if you talk about like decentralized identity like that's kind of building towards that direction and like what you do on chain is like creating this whole personal profile

um so I don't know I kind of like it too I think I think it makes sense like yeah if you tried to cheat one game and you're not gonna come cheat my game I think that's a good spot to call it uh over to the interview with Jay I cannot recommend enough for you guys to all check out blockworksresearch.com if you go over to the research Tab and toggle free research you're going to get access to some of the best free reports in the industry and if you want to subscribe to blockworks research you can do so using Xerox research 10 at checkout in order to receive 250 off and you can also sign up to our free newsletter if you want to just get a little taste where we give Alpha on governance dgen trade ideas Market commentary charts of the day Etc

kind of get you caught up to speed on everything you need to know in the Market within five ten minutes give us a follow at blockworks res blockworks res on Twitter we'll release our new reports during the week and even if you don't have access to the reports you're not a paid subscriber you can still check out the topics we're writing about and get a little bit of a brief insight into what the contents of the report is about we want know a little bit more how we think on the data side of things head over to our Dune public account we have four dashboards live there for free the revolution will not be quarterly reported so definitely check those out and let's kick it over the interview all right everyone we are joined by Jay the co-founder of say Labs uh and we are

ready to kind of get into the deep the the depths of what say is and uh you know what the innovations that uh We've really crank it out in this one and so uh Joe I'd like to toss it over you you have a pretty interesting background of kind of how you got into crypto and the things you saw that were wrong in the tradify model or maybe not necessarily wrong with things you can improve on uh with the with blockchains and 3D centralization so I'd love to get a brief little intro of uh you know what kind of brought you into crypto uh and then we kind of get into the the more technical side of things and I think what we both agree is the more fun side of the conversation awesome yeah so thanks for having me on Den Sam um and

good afternoon to you guys um so my background is that I originally studied computer science got into crypto back in 2017. um at that time my roommate he was going through binance Launchpad so we tinkered on a couple of different projects together um and then afterwards I ended up joining Robinhood right like B Robin Hood um so I spent almost four years over there I saw the company 10x and I was an engineering lead happened a little bit over two years ago so I guess for anyone in the audience who doesn't remember um what basically happened is that there was this entire meme stock kind of mania that was taking over the United States and there are some stocks like GameStop AMC there were just they were straight up moving right like they were doing exceptionally well and they were driven

by this kind of push from retail um and there were hedge funds that were shorting them these hedge funds were getting short squeezed it was a situation where retail was winning hedge funds were losing and then just completely out of the blue one day Robin Hood turned off bias they didn't tell customers why they did that no one had any idea what was happening it was just incredibly chaotic and there was National outbreaks right like there was this entire kind of like true Robinhood moment happening Robin had turned out by it and that kind of stopped that GameStop um pump so to say so that was extreme exceptionally importantly handled from Robin Hood said um they weren't doing anything malicious but it was just poor communication and as an Insider it was just really demoralizing to me there at the time because you put your reputation

on the line to join a place like Robin Hood and when they just turn off bias you get all of your friends asking you yo we just opened these GameStop positions last week what the hell is happening over here here you have nothing to be telling them and they're just kind of pissed off and they are somewhat pissed off at you as well um your team is definitely not happy and like they're asking you questions you don't know what to be telling them so after going through that experience myself I became much more of a decentralization actually because anything that happens on chin is inherently trustless and transparent so that would have helped avoid a lot of the things that we saw with Robin Hood which is why two years ago my co-founder Jeff and I we started building a decentralized exchange initially this then led to us looking into all the

infrastructure that we could use to build an exchange um and we ultimately decided that the current infrastructure is lacking which is why we started building sit so super high level say is a new layer one blockchain our mission is to build the best infrastructure for exchanges so the way that any listeners can think about it it says a layer one blockchain within order matching engine primitive that's built into it my question would be why uh Cosmos chain uh as an L1 as opposed to and also going after sector specific you guys are trying to optimize eyes for for dexes and and Order books and things of that sort so why would you go that route as opposed to um maybe if you're like a dydx going an app specific chain route

of course yeah so two good questions over there um I'll go through them first of all by talking about why did we end up using the cosmos SDK so when we initially decided that we that the infrastructure was lacking we started considering like what would good infrastructure look like and there are a couple of different three I guess three prominent uh verticals that we saw um the first was making use of a roll-up at that time it was basically building a roll-up on ethereum um the second thing was by making use of some kind of shared security layer one so this would be like polka dot or an avalanche subnet and the third approach was building a solver chain so the role the reason that roll ups didn't make sense back then and even right now in a production setting they don't really work either is because they can't scale

so with a roll up you end up having to have all chain computation and then you take uh transactions that happen on chain or often compress them and then you write them on chain so if you're writing this to ethereum L1 each of these uh transactions it ends up taking around 16 gas per bite of day data that is written to the L1 itself so currently ethereum blocks have a target of 15 million gas per block so the end result of that is that you can only get in around 6000 TPS um and the bottleneck around that scaling ends up being the data that is written itself um this isn't solved by proto-dink sharding either um with prototyping sharding you get closer to 6.7 ktps um so like realistically right now you

can't really scale if you're building um any kind of roll-up on your theorem so the second thing that we then considered was making use of shared security so keratin subnet something of that flavor um we decided that that wasn't ideal for us either because we wanted to build the best infrastructure for exchanges and one part of that is having really really good performance specifically around time to finality the lower the time to finality there is for exchanges the better and if we're making use of any kind of shared security approach we're basically Bound by whatever that validator set is and whatever their consensus time is um and we can't get anything better than that in terms of the consensus time so if for example polka dot his six second block time we can't have any time to finality faster than that so we wanted

to customize the validator set we wanted to customize the consensus that we were using and we wanted to consent just have better time to finality which is why we didn't end up making use of any kind of shared security approach um and then the third thing that we considered was a sovereign chain and even within Sovereign there's two flavors right you can either do it uh like Solana near Aptos we did which is you start off and build everything from scratch or you can make use of um something like Cosmos SDK which has already been battle tested for several years at this point um you don't need to recreate the wheel you don't need to really worry about a lot of security issues because those have already been handled um and then you can get something out very quickly and then you can basically customize this dock for whatever kind of

performance optimizations or application level changes if you want to make so from our side that made that was honestly kind of a no-brainer like rather than thinking two years into building an entirely new chain it's completely from scratch handling all the cryptography all these things that have already been done it made more sense to get started with the cosmos SDK we were able to get a working um layer one with the native order matching engine built into it in a couple of months and then afterwards we made tweaks around the consensus player adding internalization of the application layer and that was really easy to do and I mean the biggest thing about all of this is it's largely been battle tested already right like even when Tara collapsed the cosmos SDK and tenement course out of it was completely unaffected like yeah terra's algorithmic sample coin mechanism background and

tissues but the cosmos SDK I mean that was like one of the most clear Testaments to how stable it is so that's why we went down that route um to get started building sit and at this point I mean we've been working with the cosmos creators so we've been working with zucky who's our advisor we've been working with Marco who's the product lead for the cosmos SDK to push Cosmos and tender men to their absolute limits so we can get into that a little bit more in the future but yeah I think Cosmos was a really good way for us to get started so it sounds like you got a with the cosmos route because it gave you like this level of customization as a developer that you felt like was a great starting point to then build further out into but it also has this

like proven track record uh of being a very comprehensive and capable blockchain Network and that's like the Brilliance of the cosmos SDK as well as tenderbit uh but yeah continue yeah yeah I mean and to that point I think this ability aspect of it is actually something that people don't talk about as much like if you look at Solana for example the reason Solana has outages is not because they have some fundamental flaw in their mechanism I think from a theoretical standpoint Solana is extremely well thought out um I think the issue that they run into is they just built everything from scratch and then there ends up being application Level bugs or like consensus level bugs which then run into teachers so there's just a lower chance that there being bugs if you're using something that is already this stable um so that was definitely one of the

reasons that we chose to go with Cosmos um with regards to the second question that you guys had asked around sector specifications um the high level way that we think about it is that there's essentially two uh kind of extremes around uh layer one infrastructure out there right now on one hand you have general purpose change uh like ethereum and Solana and on the other hand you have application specific ones like osmosis and dydx before um we think both of these sites have pretty significant benefits right if you look at application specific chains the core idea over there is you have one application um and then you customize the entire stack so the application Level you customize the consensus level you can customize the Storage level like every single part of it around that one application so by definition you are

likely going to be building the best possible infrastructure for that specific application because it's completely built up around that application um that's a really powerful thing because you can have much better user experiences you can have much better uh like performance and you can make whatever optimizations you need to so that's something that we really like about the application specific approach um with the general purpose approach you don't get that level of customizability but what you do get is social coordination so every single project that is building in that ecosystem they're all on the same team and they're all working together to help that ecosystem succeed and that's why if you look at the biggest essentially the biggest projects out there they all have this general purpose approach because it's much easier to scale and grow an ecosystem if it's like let's say dozens of different teams that are

collaborating then if it's just one application that is basically streaming into an abyss trying to get users trying to grow the community so because of that there's kind of two takeaways the first is an opt-in approach is going to lead to better infrastructure but it's much more difficult to have it grow and scale from the community side and we think the community side ends up being one of the most important things especially when you have like open source technology like the technological mode ends up being pretty limited um the community ends up being the biggest thing uh that ends up leading to long-term sustainability for a project right like if you Fork ethereum you're not going to be able to get the same kind of success as ethereum like because ethereum has this community already so um the approach that we ended up taking was kind of a middle ground approach

which we termed sector specific so it benefits from the idea of having customized infrastructure so we basically take in um the cosmos SDK and built infrastructure that is customized for exchanges through changing the consensus level application Level and also adding in a native order matching engine um and the other thing that we've done is we're not building an application ourselves so we're just building the core based infrastructure and then we're allowing other applications to come and build on sale so be good to that a couple of things one is that we can now have an ecosystem of applications rather than just having an option um and the second thing is now there's a sense of credible neutrality so if you're building a desktop chain and then a DEX wants to come and deploy a smart contract on you

um that Dex is going to be competing against the core app of that option so from that uh basically that other exchange standpoint it doesn't really make sense strategically to be deploying on that layer one because you're going to be competing against the core infrastructure that you're building on so with the sector-specific approach we get credible neutrality we get social coordination and we also get the customizability that I was talking about before so we think that ends up being the best approach specifically for these types of applications that have product Market fit like exchanges and then need some mechanism to help them scale and have better user experience awesome yeah and I like how you really focused on the the the fact that communities are emotes because you know again we see that over and over again throughout the industry but I want to get a little deeper onto

the technical side of things so you took the cosmos SDK and tendermint as a great starting point but then you made a number of technological breakthroughs that uh I think as you put it we're pushing uh the stack to its limits and so I'm curious where the Divergence comes in right so you know I I've seen about the the twin turbo consensus optimistic block processing parallelization um there's a there's a lot of different things to go through here but I was like I want to post it to you like how do you however you want to break these down like what are the most exciting technological innovations uh that say was able to push through onto the cosmos SDK and tendermint yeah absolutely so at a high level we're building the best infrastructure for exchanges the question then becomes what does that infrastructure look like

um so from our side the way that we answered that question was we went and talked to different exchanges and asked them for feedback around like what is actually important to them um after having enough conversations there were basically two verticals that became clear that these are important exchanges right the first is performance specifically around throughput and latency um and the second is around user experience and that's much more of like a vague thing um that is the different exchanges have different ideas for what user experience looks like but everyone wants to have this idea of better user experience so we started initially um building by adding in this native order matching engine which I can talk about a little bit later and then afterwards we thought to ourselves how can we really push the performance aspects of this um to be even better so that's when we

started looking into the consensus layers and the application layers and there were basically two things we started doing to enable better latencies and better throughputs so from the latency side we added in Twin Turbo consensus so this is taking tender mint and we changed the way that block propagation Works uh we changed the way the block processing works and this allowed us to have tender mints that was running as fast as theoretically could um and the second thing that we did is we added imperialization so we're currently the only Cosmos chain to be making use of any form of parallelization so I mean Super High Level D kind of result of these changes were that we are currently seeing a 500 millisecond time to finality in our latest public test net which makes us the fastest chain to finality out there Point Blank and this is with conditions

that may make midnight um and the second thing is we've been able to get in that public test net up to 20 000 orders per second that we can process so most ecosystems have around one to three thousand orders that they can process every second so roughly a magnitude better over there as well um in terms of more of a deep dive into this actual like twin turbo consensus parallelization um twin turbo consensus is basically taking tendermen and then the optimistic block processing and intelligent block propagation were the two things that we did which is why we called it twin turbo consensus um so I can start off by talking about block processing um the idea of optimistic block processing is honestly pretty simple if you look at temperament the way that it works is first it'll be a block producer that proposes a Blog then validators

will receive that then there will be a pre-vote step where they send messages to every other validator then there'll be a pre-commit stuff where they once again send messages to every other validator and then afterwards they'll actually execute the transactions in the blog um and then afterwards they'll commit them so execution typically with tender mint and Cosmos SDK happens after the two voting steps are done now if you think about it this is really really inefficient because validators received this blog and they're just chilling right they're not doing anything for the pre-vote pre-commit set and then they only start execution after those are done and we thought to ourselves like we can definitely optimize this because like the core thing over here is you can start having a concurrent process spin up as soon as the validator receives the

block and just have it update some kind of candidates did and that's exactly what optimistic block processing is um a validator will receive the block and then they'll start executing immediately um and they'll be updating the candidate State and then one of two things will happen after the pre-commits episode either the block will be approved in which case you can take that candidate State as a validator and then you can just commit that and have that be the canonical state of the chain so in that case it's really simple and you end up having much better performance um the other scenario would be if that block is rejected by the network in that case you can just discard the candidate State and then you can start moving on to the next log and the the core concern over here is that what if there's malicious validators like let's say 30

of the validator simulations and then they somehow try to DDOS the validators the network by having them just try to run a bunch of optimistic blocks um like one after the other so the way that we got around that is we only have the first block for any given height the optimistically processed and after that block builds and every other block for that specific height will be will be processed sequentially like after the pre-commit step is done so this helps prevent like this ddosing scenario where like all the resources just end up getting used because their validator is trying to have like 10 different or like I mean notes even more than that um concurrent optimistic processes that they're trying to run um and any questions around that piece yeah so I do so I'm assuming does this affect block times at all so if the the

block is optimistically uh created then I'm assuming there'd be a faster process right and so that's probably where you get the 500 milliseconds but what happens if it's rejected and it goes to like defaults to sequential ordering yeah so so a couple of things over there the first is yes in cases where there's higher throughput this will help reduce the block time um in terms of the lower balance this is not really going to matter because in lower bound there's actually nothing to process so that ends up being more or less negligible um process with that but if like the actual execution part of it is taking let's say 500 milliseconds and then each voting step is taking let's say 100 milliseconds then those two voting steps that take 200 milliseconds you basically have 200 milliseconds worth of that

execution happening during the concurrent processing step so it definitely helps uh in these higher throughput scenario times um with regards to your question around failure cases so if it's rejected by the network then what happens is we will move on to the next block to that height and that block will not be uh I guess processed using this optimistic approach and that's to prevent that default scenario that I was talking about before so only the first block at any given height will be optimistically processed and then afterwards from the next height onwards you'll start optimistically processing again so this is is kind of something that restarts every single um at every single height no I'm curious too like in comparison to a regular Cosmos tendermint chain how

does this affect Hardware requirements considering the performance is so much uh better yeah so optimistic block processing is actually not something that will really result in additional Hardware requirements um because you're basically taking something that'll be happening after the execution set or after pre-commit and just running it as soon as the block is received so there are other things to reduce specifically parallelization that end up being more resource intensive than uh for example like what you would need with like ethereum um but the optimistic block processing is something that other Cosmos chains can do as well without increasing their Hardware requirements and that's something we're actually chatting with um the teams behind this as well like we're probably going to be upstreaming these changes and then having them deep lives that teams can use if they want to

um because at the end of the day this is all open source anyway like other teams can make use of this if they want to so we'll just make it easier for other teams to benefit from the kind of innovation that we've done so you mentioned parallelization there and you know this you hear this term thrown around a lot like I even hear it with polygon they're working on that for their previous day commit chain so can you kind of explain the nitty-gritty around parallelization and the benefits it uh it provides yeah yeah um I guess before I get into that I'll hop into intelligent block propagation which is the other participant turbo consensus and then afterwards we can do a parallelization deep that's cool yeah so from intelligent blog propagation standpoint this was the other core part of twin turbo consensus and I mean basically what happens right

now with normal tendermen is that a block producer let's say that you're a user you want to submit a transaction the life cycle life cycle of that transaction will be first it goes through a full node and then it ends up in the mempools for different validators um then one of these validators will be a block producer so they will look at their local mempool and they'll construct a block based off of that once they have that entire block constructed they'll first create a block proposal message which just has a block hash and then afterwards you'll send that block proposal message over to the rest of the nodes in the network then they'll take that block that they have they'll break it up into a bunch of small chunks and then they'll send all those chunks over the network so any validator that wants to start

voting as part of their pre-vote step they'll need to wait to First receive that blocked proposal message then they will need to receive all the separate blockchains they'll need to put all the chunks together find the hash of that block make sure it matches the hash from the Block proposal message and then they can proceed with voting right now the thing that's extremely inefficient about this is that the contents of the block that these validators have to wait for it's just all transactions and most of these transactions generally all of these transactions end up already being there in the mempools for each of these validators so you're basically waiting to receive this data that you already have which is just incredibly inefficient so the approach that we started taking instead is that log proposal message that I described before

instead of just containing the block hash it also contains transaction hashes so it contains an ordered list of each of the transaction hashes and the transaction hashes are much smaller than the actual transaction so the result of this is the block proposal message will be a little bit bigger but it's not like massive and you can still send it as one message to every single validator in the network then the validators will receive this block proposal message they will look at their local mempool and if they have all of those transactions because there's already a mapping of transaction hash to transaction in the nipple so they can just look at uh to see if they have all these transactions if they do then they can just construct the block locally so the massive benefit there is previously they were waiting to receive all of these chunks over the network and

there's costs associated with this network vacancy right like it might result in 100 milliseconds if additional time it takes to receive that um now they don't need to wait that anymore they can just construct it locally so in the happy path they can construct it locally they can begin voting and in our case optimistic block processing immediately after that so that's a huge win um the question then becomes what happens if they don't have every single transaction from that block proposal message in their mempool um in that case we just fall back to using what we were previously using which is the block proposal will create this bigger block proposal message that has the ordered list of transaction hashes and then they'll still take the blog break it up into chunks and send it over the network so if you as a validator you do not have all the

transactions from that block proposal message then you can just wait and you'll eventually receive the transactions uh from the form of that block and then you can just go ahead and use that so in the happy path it'll lead to substantially improved performance and in the case where you don't have all the transaction hashes you can just wait a little bit longer and then you'll essentially fall back to the same time um the same latency as you were seeing uh in the normal tenement approach so this is really exciting it almost feels like the approach you took and tell me tell me if this is completely off base but it was like we see opportunities to improve the efficiency of a few of these systems and so we're going to build these systems but there are case edge cases where they're not going to work 100 of the time I mean in

the event that happens then let's just use this already great infrastructure we're building on top of and rely on the ways it's been running in the past is that something you'd agree with exactly yeah so we're very optimistic in our approach like if things are under ideal scenarios we'll have better performance and if things fall back to the kind of normal scenarios and we'll just use what's already been battle tested with Cosmos and tendermint and fall back to using that already reliable approach now that's pretty exciting and so you guys are already in uh public testnet and when you look at something like the the you know you look at the data of what's been happening thus far how often are is like uh intelligent block propagation or optimistic um uh processing how often are those uh being enabled how often are they being

utilized within this test net or how compared to falling back on the traditional methods yeah I mean candidly for this public test that we haven't collected metrics on that um in the internal test net for intelligent block propagation it was 99.99 of the time that that was applicable um so yeah I mean we probably should con uh get more metrics on this in the public test net but um the block times just given that they are as low as they are right now um even under higher throughput scenarios we do believe that it's getting applied basically all the time and like the core thing over here is if you have a smaller validator set then it's much more likely for these to be beneficial especially intelligent block propagation as soon as you start having like hundreds of thousands of validators or something then the intelligent block

propagation approach might not really be that helpful but if you can make the assumption that you have like connections between every single validator which you essentially need to have with the tenement like if you're using tendermint um then you're able to benefit from those assumptions and have better performance around that that makes tons of sense to me let's dive into parallelization this is this is an interesting one yes sir so parallelization is for any anyone who's listening that isn't familiar um the normal way that like ethereum example is processing transactions is let's say there's 100 transactions it'll sequentially process them so go through every single transaction one by one and it's not able to do any of them at the same time now this is really easy easy logically um it also results in lower Hardware requirements because you don't

need to have any beefy machines to be able to um better process transactions the downside here is it is really slow um and it's not very scalable which is why on ethereum for example you can't really have super scalable applications that are built on the layer one itself just because they can't really handle substantial throughput so the approach in web 2 that was taken to help combat sequential processing is this idea of parallelization which basically means that rather than executing one thing at any given time you can execute multiple things and as long as they're not conflicting things that are being processed you can have separate work streams that are processing each of these and it'll allow you to finish processing these workloads in a faster period of time right so the core idea is something that is applicable to

everything that we use um like basically any anything you do in web 2 you're going to be benefiting from some form of power parallelization to have better performance um we decided that we should do that on say as well so we're currently the only Cosmos chain to be making use of any form of parallelization and the approach that we're taking is that uh developers will need to first of all deploy their smart contracts then afterwards they'll need to uh execute a transaction that'll Define the dependencies for each of the functions in those smart contracts so it'll basically be a template that they need to Define and then at runtime that template would get filled in um and then afterwards when at runtime transactions are being processed uh you'll be able to look at the

dependencies that are being basically the state that those transactions are touching and then you can correct construct a directed acyclic graph of these dependencies so an access list of these dependencies and then anything that is touching the same state so if both of us are trying to touch the same order book then our orders would need to be processed sequentially but if we're touching separate order books for example or if we're touching completely separate things um then our orders can be processed in parallel so if there are paralyzable workloads which basically means workloads that are not all touching the same shared state so easiest way to think about that would be like different people sending tokens between each other are different people touching different smart contracts um then parallelization helps significantly improving um the amount of transactions that you

can process and if it's some if it's a workload where like everything is touching the same state then in that case parallelization ends up not being as beneficial because you still need to be sequentially ordering or executing those transactions so uh the approach that we've taken has helped us get up to around 20 000 orders per second that we can process in the public test net so it is quite beneficial from a performance standpoint um and I would say that the approach that we're taking there's basically a spectrum of approaches to parallelization as well um on one hand you have apptos with their optimistic processing approach and then on the other hand you have like Solana and tweet where every single transaction uh you need to include all the accounts that are being touched or all this data that is being touched um and rather than taking either of

those approaches we're honestly taking a middle ground approach where developers just need to Define these dependencies one since submit this transaction once and then afterwards for every single incoming transaction they don't need to Define which accounts are going to be touched or which state is going to be touched um the chain is able to decipher that at runtime that's pretty cool and so when I think about this we recently had John Adler on the podcast he's building uh Fuel and he talked at length about parallelization and one thing that helped it kind of Click for me was like okay you know like I'm coming at you from a laptop with uh with an i7 processor seven cores in there and then like on the number of course increases which allows them like aerialization basically allows the course to be working at the same time that kind of helped it click for me uh

just for the listeners out there as well and so you know like it's just this hits the point where it's like okay so now we have all this improved Tech um but like you start to ask the question like where are the trade-offs here uh because you know as we see with a ton of things in this industry there seems to be uh everything seems to be on a spectrum with you know and we have trade-offs everywhere so where would you define the trade-offs uh with a chain like say yeah that's a great question and I agree that everything ends up having fatals like anyone that says that they've solved every single problem around the scalability or around the um I guess trilemma they're probably uh exaggerating so in our sense uh the biggest thing that we decided to do is we made use of Cosmos and Thundermans

get started um and tendermen specifically is the biggest feared off that we made because tenement scale is N squared which means and the reason for that is in order to basically create a new block you need to have two-thirds consensus between all the validators so essentially all the value validators need to talk to each other before you can add a new block to the network um what this means is that you can't actually have a lot of validators um if you're trying to reach consensus quickly so that's why Cosmos help for example is an example of the tenement chain with the greatest number of validators guess how many validators it has 150. yeah I think it has 175 so roughly roughly in that range right um so you can't really get past 200 validators I don't think there's an

example of a cosmos Vision that is that is big and that has done that um and as a result of that like we expected that we're going to be capped on the number of validators unless we make changes to the tender midwich and honestly we do have plans for changes that we can make to enable a larger number of validators um but assuming we get started with like some vanilla implementation of tenement um you're not going to be able to have a large number of validators so that's the biggest trade-off that we made um and we decided to lean into that because if you if you make the assumption that you have N squared connections and every validator can talk to every other validator then you're able to make performance optimizations that benefit from that one example of that would be twin turbo consensus um like specifically the intelligent

block propagation piece if you assume that all the validators are stable to talk to each other and they do end up having transactions in their mempool um then you're able to have better performance around that whereas if you're assuming that it's like 100 000 validators a transaction being in someone else's mempool is unlikely then those performance optimizations we made wouldn't really be applicable anymore so yeah I would say the biggest trade-off was using tenerment and I mean we do have ideas for how we can we tender mint in the future to enable a larger validator set size so staying on the topic of the validator set what is the size of the validator set in the public test set today and how many do you plan to go live with at launch and not to throw too many questions at you at once but like what

was the selection process there was that a white listed uh validator set or open to the public curious about how your approach was on the value I said specifically yeah so I believe it's 35 that is the target cap right now for the public test net um it'll be something in that similar range for minutes like somewhere in the 30 to 50 validator range um in terms of like when we got started we were thinking holy it's gonna be super hard to get validators like there's all these different projects that are trying to offer shared security like the core problem they're solving is like it's hard to have like a good validator set um it turned out to be pretty easy from our personal experience like we and I think that's also because our community was like growing pretty quickly so there was like at that time like when we got

started with our uh incentivize us that there was probably a pretty high degree of hype around the same project like more than we had originally just did it so we ended up having over 2 000 validators that signed up or that wanted to like submit an application to be part of the validator set um and then from there I mean the way that today's validator set works is it's completely permissionless so anyone can join if they have enough delegations um what state was doing initially for the incentivized testnet is we picked 200 validators to be receiving valid that'll be receiving delegations and then we just alternated uh the delegation to them so that everyone got a turn and then afterwards the validators that were like really strong performers from that incentivized

cessnet um are now the validators on Atlantic too but it's completely permissionless so anyone can join if they get a high enough delegation and on mainnet it'll be somewhere like um on Main at the same Foundation we'll be delegating to some validators that are strong validators but anyone can join that validator subside or that validator said if they have enough delegations and so how will gas costs be calculated and what will they be paid in yeah so gaps will be paid in say tokens um in the future we might allow for other types of tokens but right now it's simplest to just start off with day tokens um and in terms of the costs around that we still need to fine-tune those settings but it's going to be extremely cheap like most likely less than a penny for every single transaction and insane particular like the native order

matching engine does allow for market makers to submit batch transactions so the biggest party that would need to be submitting a lot of transaction to the market papers and they basically only need to submit one transaction that is composed of all of their orders within that transaction so we don't really expect anyone to be paying substantial um gas costs in the future so the token conversation is always fun to save for the end but now that we kind of kind of got towards that direction I'm curious is the state token going to look something similar to like an atom or an osmo uh like a standard proof of stake uh proof of State gas token or or whatever the any additional utility around the token yeah so it's going to be more or less a standard proof of State gas token so similar to Solano um it'll be used for

three primary things which will be governance security transaction fees um the other thing that's really interesting about say is say it's a trading Focus ecosystem right so there won't be realistically a lot more Mev as a percentage of total activity on say and that would be in another ecosystem like maybe for example Solana that doesn't have as much trading happening compared to what we envision would happen would say um so as a result of that we think that Mev redistribution will be another mechanism for Value pool um and the way that I mean I guess high level for uh listeners so Med stands for maximal extractable Value it's the value you can extract as a block producer by changing the ordering of transactions or

including or excluding transactions within a block um and there's this idea that um you can't really stop Mev altogether so rather than trying to solve it you can have auction so that in the case of ethereum with flashbots it's the validators who receive the money by having people bid to be included in iblock um and we're taking this idea and we're taking it one step further which is we're going to be having auctions happening where people can submit transactions and then whoever wins these auctions they'll be included in the block um and the thing that we're doing that's different though is that whoever wins these transactions and then submits these orders in the block these bids will then be taken and then redistributed to validators and stickers of the chain so through an approach like this it results in

um much more avenues for Value accrual because any of these going to essentially like the value of Meb in the future is going to match whatever this or the value of Mev redistribution will match the side of that and maybe opportunity like if it's 100 opportunity and there's 100 different people that identify that opportunity they will all competitively bid up the bid that they're making until the bid is as close to that 100 opportunity um as it can be so realistically to be like 99.50 or something for the bid and the value of that bid will then get redistributed to the network so uh we anticipate that to be a pretty interesting source of value accrual in the future um but that's not something that has been built yet that'll be built after a minute launch that's super interesting I was just

going to ask you so how do you expect to attract liquidity because obviously you need a lot of liquidity in order to increase the amount of Mev that's generated but considering the staking yield will likely be higher considering all the trades being made have a lot of high Mev that that's an incentive in itself but is there any other strategies that you have planned there yeah I mean the distort there's going to be or drop incentivized test and rewards so that'll be a chance for folks to get um basically start participating which say pretty early on from managenesis um beyond that I mean we're going to have all the infrastructure set up to enable people to bridge stuff over to say so IBC from the get-go will have different bridges that are set up as well um but beyond that we think it's going

to end up being kind of a gradual process to start having more and more activity options I think originally it'll be the degens from like other ecosystems like ethereum Solana Cosmos dgms to come over um and there will be applications that are built on save from day one so they'll be able to partake in those applications these applications will realistically be offering some kind of incentive for people that are trading so we think initially it'll be the degens that come on after that we think that there will be more retail-focused applications that are able to get CMO and say um and then that's the biggest bet that we're making like if there's no like killer opposite then say it's an ecosystem is probably not going to be a massive ecosystem um but if there is a killer app that is able to take all within the next six

months to two years um then there will be a lot of retail folks that come on as well and I mean we've seen this happening in the past several years where any kind of killer application in the ecosystem results in a lot of retail coming on um and then from there we anticipate that there will be a lot more liquidity coming in and as soon as their Sweet Child liquidity then there ends at the institutions that become much more incentivized to participate as well so we think the third group that'll get involved on is institutional players so um from our side the biggest thing that we're focusing on is just doing it's as good of a job as we can of building the core based infrastructure and then chatting with teams kind of explaining the benefits of building on state to them and at this point we have over 120

projects that are building onset so I think that overall there has been a lot of interest from different projects around initially it was just the core infrastructure like being able to build an exchange that is potentially much better than an exchange you can build in any other ecosystem um and now I think it's just becoming more and more strategic for teams to launch onto it because I mean exchanges are one of those things that become kind of magnets for other types of applications to build around um this is true both off chain and on chain like yeah I mean we can talk about like how binance has a similar thing happening how ethereum has a similar thing happening but in any ecosystem where there's exchanges other types of applications get built as well so we're starting to see more applications just come um and try to strategically benefit all

these exchanges right on and just going back to quickly to the MVD discussion so you know there's like good Mev and bad Mev right so like an example of bad MVPs like front running or sandwich attacks where you're throwing off the price uh that a user is getting on a trade but there's also a good mov uh in the form of like liquidations and whatnot so say it has an interesting front running protection mechanism can you walk us through how that's built and the benefits that provides to its users yes absolutely so I guess the other core technical Innovation that she has decides the twin turbo consensus and parallelization is a native order matching engine so this is something that's built into the chain itself and anyone that wants to build an order book based exchange um can make use of this underlying order

matching engine to help with every step of order processing so or replacement order matching order settlement order cancellations um that'll all be handled by this order matching engine and what it actually does that's really interesting is it makes use of frequent batch auctions so during order processing it'll aggregate every single order together and then at the end of the block it will execute all these same all these orders at the same uniform clearing price um so one example would be if there's an order book for let's say um ether right now and the price of ether is what is it it's 17.91. so if there's two orders on the order book one to sell either for 17.90 one to sell either for 17.92 um and then two Market buys come in so with sequential execution which happens

on other chains the first buy will get filled at 17.90. the second buy will get filled at 17.92. so that means the order of transactions within the block does matter whoever in that case is first they'll get a better price um in today's case both users would get the average price which would be 17.91 right 1790 plus 1792 divided by two um so there's kind of a couple of things that happen because of this first of all there's price fairness within a Blog which just leads to users being happier because otherwise it is kind of annoying to see other people getting different prices than you within the scope of a block um the second thing is that front running is no longer possible for any transactions related to the order book that are included in the block because the way that front running would typically work if you have Atomic front

running would be someone sees my order to buy either for 17.91 plus like let's say I accept one percent slippage so 17 uh that would basically go up to 1791 plus roughly 18 so up to 180 uh 1809. um so what you would do in that case Dan you would buy ether for 17.91 and then you would just have one transaction composed of multiple steps where you buy ether for 17.91 then afterwards you place to sell a limit sell to sell it for 1809 and then assuming there's no other liquidity on the order book I would end up purchasing it from you for 1809. so in a riskless way you made that one percent slippage is profit for yourself and I just kind of got screwed from that because I paid one percent higher fees when

um it's basically I I didn't need to pay that it was extremely exploitative right so in traditional financial markets that's legal um but in the case of say because we're making use of the frequent batch auction approach that's not even possible to do so that kind of ties into our overarching approach to Med which is we want to stop Mev we want to stop Adam UV like frontrunning and then for other types of neutral maybe such as liquidations arbs kind of team ins um I mean first of all we don't really think we can stop those and secondly we don't really see the benefit in solving them if they're not negative for anyone so we'll let sophisticated players benefit from having opportunities to um basically participate in these types of movie and then we'll just have most of the value of those opportunities get redistributed to the chain through the

mechanisms I was describing earlier no that that makes a ton of sense uh and even just thinking about this through my in my head now I kind of like stacking it on the previous discussion right so how does parallelization work when you have like multi-hop trades right so say I was going like an eth to bitcoin like I wanted to go East to usdc but for whatever reason that pool didn't exist so I had to go like East to bitcoin to usdc how does parallelization work and or like even the order batching in that like I feel like you know if I went eth to bitcoin to usdc and then for whatever reason I want to go back to eth all in the same trade like would I end up with the same amount of heat there is uh like you like walking through how that works

yeah so I mean there would be two types of ways that this could happen one would be through normal smart contracts on today so like an amm contract or an order book contract that's not making use of the order matching machine um in that case it'd be identical to what would happen on ethereum where all those would be supported um and the end result of that would be like if you go through that entire cycle you might not necessarily get that same price again um the the other type is for exclusively order book related um transactions that are tied to order books that are built using this native order matching engine um in that case you can Define dependencies so you could technically like whoever made the markets for you said ether to usdc or either to bitcoin and then Bitcoin to uscc um they could Define dependencies

between those two markets and then one of them would run first all the time and then the second one will always get run afterwards in that case you could have that multi-hop um trade workout but realistically you're not going to have dependencies being defined between like every single uh Market out there so in those cases you will not be able to have this multi-holp trade happening because the frequential approach that's a trade-off that happens to that approach because you're having everything get executed at the end of the block and there's only one time that those trades get executed having something that like goes multiple Ops across different contracts um that wouldn't necessarily be possible to do so yeah like what we intimidate happening is uh sophisticated traders that are trying to do these multi-hop approaches generally for collecting art of some kind they'll have to take on

additional risks to do stuff across multiple blocks um and we think that's totally fine like the main thing we want to focus on is improving the retail user experience and if there's retail users in any ecosystem then sophisticated Traders will start uh kind of partaking in that ecosystem even if things are a little bit uh more difficult for them and if there's no retail users and like sophisticated Traders don't really want to get involved anyway so that's why it makes more sense to enable the best experience for retail users yeah that makes a lot of sense now in terms of pricing that's obviously a very important factor for any decks can you go into kind of Oracle solutions that you guys have thought about or are you guys leaving that to uh the actual dapps building on top

yeah so from safe side we're supporting a native price Oracle um so this is going to be something that is uh built into the chain itself and every single validator will need to submit price feeds every single block for some small list of governance approved assets so this would be extremely high liquidity assets like Bitcoin and ether um and in that case it'll be extremely decentralized because it's extremely trustworthy as well because it's literally the validator said there's securing the chain that is providing these price feeds so as any kind of exchange that wants these price feeds um you'll have probably one of the best data sources available for that um for other types of applications so maybe you want prediction markets and you want like the outcome of attendance

match or maybe you want to offer like perps on some asset that is like kind of a new obscure asset that's not uh that the native price Oracle on say it's not providing a price feed for it um in that case we already have multiple projects that are going to be deploying oracles on set so a lot of the bigger names um are going to be deploying or so you will be able to make use of those prior quotes as well um to get those kind of long tail asset price feeds or to get those one-off kind of prediction Market related price feeds as well so this part is fascinating to me because the cosmos doesn't really have like the the chain link of its ecosystem yet uh do you think that these price feeds could really be propagated throughout the cosmos ecosystem as a

whole they definitely could be yeah so I mean because of IBC it's pretty easy to take a transaction that happens on sale and then have the results of that be propagated to other chains um the downside over there would be that they'll be somewhat still like something that gets written to say at times at time T um it'll take essentially six to seven seconds to make its way to another chain right because it'll require whatever the relayer latency is plus afterwards it'll need to get included in a transaction that is part of a block on another chain and go through consensus so it will be still prices so if you want prices if you're like using some kind of DeWalt um in that case it should be fine um to have like a little if you're using like a 30 minute too often having like to

wait seven seconds for each price Suite that's that's not that big of a deal um if you're not using a t-wop and you're looking at like literally the latest prices I think in that case it might not be ideal to have this kind of like IBC setup from CSI yeah so that makes a ton of sense and when I think about the price feeds right so validators are pulling in external price data to then um you know propagate throughout the network and allow other applications built on say to benefit from and so of course the validators themselves aren't financially incentivized to act in the best interest of the train as with most peripheral State chains uh but how do you promote the decentralization of the external price feeds that these validators are pulling in uh to then you know submit throughout the chain yeah so from kind of a game theory

standpoint we're not really doing anything from suicide to facilitate that um in terms of the like say Foundation realistically we'll be able to get involved and encourage validators like we'll only give delegations if you're providing price feeds from these separate sources um but from more of like a chain level Game Theory standpoint that is not something that is important to say um and for what it's worth like there have been examples of other projects where that ended up being perfectly fine like in Terrorist case for example it was something similar where there wasn't really any game theory mechanic that enforced different sources for their data um and that ended up being fine as well so I think for most applications they're not really going to get affected by that and if a certain application really does care about having like a lot of different sources for their um

whatever price feeds that they're looking at they can make use of any of the Oracle solutions that are going to be deployed on today as well so for those specific types of applications there will be Solutions as well awesome and I want to dive a little into the ecosystem as well obviously we're in public test right now but um you know I heard that there was over 120 teams committed to building projects on say um not so much about the brand itself but which Pro like projects or like types of projects are you most excited about you know I I understand the State team itself is not building any of these projects to kind of keep that credible neutrality um but you know you spend a ton of time with that with the technology and likely know it better than most so I'd love to

get uh your insights onto what you're most excited about coming to say yeah so because of that native order matching engine that we have built into chain itself um the types of applications that directly benefit the most will be anyone building an order book based exchange they can either build some kind of middleware or build some kind of exchange that leverages this uh Native order matching engine um one example of a project like that would be cargo so cargo is building prediction markets middleware where they're just building the core kind of infrastructure and then other people can build each of the exchanges and they're leveraging the native order matching engine and then they're allowing each exchange to be geographically specialized so rather than them supporting India and Europe and like uh the United States for different types of prediction markets they're having different teams that are geographically

specialized um that are building on top of that so in general I do think that ends up being the native order matching engine ends up being one of the more interesting use cases um besides that like anything that benefits from a lower time to finality does benefit from building onset so having this 500 millisecond time to finality results in a vastly improved experience for any type of I mean honestly any type of exchange like if you the market maker are trying to open a position on chain and then have it get hedged on off chain um typically we'll wait until after that trade has actually been finalized so that it doesn't get pre-ordered away in that case if it's like a 15 second time to finality um and you're only going to be hedging it after the trade is finalized um then you're essentially taking on

that 15 seconds of volatility in 15 seconds at risk which results in like the way that market makers basically account for risk is by offering wider spreads um which then results in a worse trading experience for users so I would say anything that benefits from lower time to finalities would definitely benefits as well um and I think that that's honestly been resonating a lot with teams at this point we have over 120 projects that are building onto ahead of the minute launch so when we originally started announcing our first test net this was around May of last year I call um that time I think there were around 10 projects that we're building on today and then afterwards there's been a lot of kind of activity happening in the crypto ecosystem there was Terra there was FTX there was um I guess even even in the past couple

of weeks there was drama as well so we started seeing more and more teams become interested in building onset so yeah I mean at this point um 120 projects I think on the ecosystem page and on social media over 60 have already announced so if you want to kind of look through those you could go to insane network.io website and yeah we're gearing up for a mainnet launch coming up pretty soon now so probably sometime Q2 in the uh around May to do in time frame of this year um and yeah we're pretty stoked everything's been going pretty well for the public test net so yeah thanks for finally coming together all right so I'm curious um for airdrop mechanisms I know I saw you tweet the other day that you were asking for you know the best mechanisms that you people have seen

throughout the space so uh are you thinking of doing something novel or the typical model of you know give Adam stakers a little piece and then uh LPS that kind of thing yeah so I mean from Seaside we really will be doing something that is broader in scope than just giving it to atom stickers um I think that some of the mechanisms that like I guess that there's fundamentally two ways of doing their drop one is like a preemptive air drop and the second is kind of a retroactive airdrop that we've seen so far um for a preemptive airdrop I think you don't really know how people are going to be participating the end objective is to get as many people to get exposure to say as possible so I think one of the things we're definitely going to be exploring is having people um receive an airdrop like you'll be

eligible for some percentage of beer drop right away and then afterwards you have to actually do stuff on chin and like try out the product and like participate in the community and ecosystem um to unlock the rest of it so I think that's a much better approach than like giving 100 of the air drop out to users right away because then they're just going to dump and they might not ever use the ecosystem um beyond that I I do think giving retroactive airdrops is a really interesting approach the only problem with like an L1 blockchain is you need to get tokens into the hand of the people more or less immediately after Genesis so a retroactive airdrop like that doesn't work as well um for an L1 blockchain but with that being said there's like other ways you can incentivize user Behavior as well and then kind of forward them

retroactively so those are some of the things that we're investigating from our side the only thing that we've kind of publicly announced that first of all we'll be having an incentive by assessment uh like incentive assessment is going on right now one percent of the total token Supply will be given out as rewards for the incentivized assessment so anyone that's listing right now that is not played around with his incentivized assessment um you should definitely check it out it's really easy to get started and you don't need to be like a validator to participate you could just go ahead and trade on one of the apps that are built onset and then you'll be eligible for receiving the inset devices Center words um the other thing that we've mentioned is that there will be an airdrop percentage of the airdrop is still TBD but there will be an air drop and then

folks that receive the airdrop they'll be able to unlock the tokens and then start uh using them on say ideally for mainogenesis and one percent for the incentivized test net I got to get on there and try it out yeah we really want to yeah we're really into invited people through that I I don't think most projects are about incentive by assessment that's not large so um yeah definitely I think that's been one of the reasons why sage has been able to get more kind of action happening in the testament which has been super helpful for like testing out the chain right like otherwise it's really hard to get mainnet type of traffic and like people trying out different kind of spam scenarios and different attack vectors um so I think the intense by assessment has been pretty helpful for getting both the word out about saying also people to

just try it out yeah makes a ton of sense now I don't want to end on a token related question because that's too cliche so I got one more question for you I guess it just makes a lot of sense for me for you guys to be working closely with centralized exchanges in lieu of all the FTX stuff that we saw this past year and one of my biggest frustrations with the cosmos ecosystem is you know sending from an exchange to the hub to like wherever you want to go next so are there any plans to uh partner with an exchange either to actually use sales Tech or to at least support uh say bridging directly from the exchange uh so for centralization listings yes that will be set up by minute so it should be a much cleaner user experience than having to go through multiple Ops

to be able to get tokens on set um yeah I mean that's honestly really important for having a good user experience so that will be ready uh the second thing around centralized exchanges building on say like building a tax on sale that's interesting because we actually have had conversations with exchanges around that um nothing has been finalized so I can't really talk about that right now but it is pretty clear that like centralized exchanges they do see the value proposition of also having a decentralized component or some kind of hybrid component so there is actually pretty strong interest for them to um I guess right now they're exploring different ways that they can do stuff on chain like some of the bigger exchanges they might just build their own layer one um but a lot of other exchanges they're interested in deploying projects deploying exchanges on chain and then

having that tie into their core centralization somehow um so I actually do think that within the next like year year and a half there will be at least one centralized exchange um that launches something pretty meaningful I'm saying that's exciting stuff awesome Jay well I'll hand it over to you to tell people where they can find you learn more but thank you so much for coming on we'll have to do this again uh once you guys are fully ramped up yeah yeah thank you for having me on Sam and yeah I mean for anyone who wants to learn more about me about say my Twitter handle is jandrajoke it uh it's just at gender joke so j-a-y-e-n-d-r-a j-o-g um on Twitter uh first day Network it's s-e-i-n-e-t-w-o-r-k um I would highly recommend following CA Network Twitter and then from there you

can see the link tree and then access like Discord learn more about instant device testnet um and participate in the Atlantis program so for anyone that doesn't know the Atlantis program is a fully community-led program that has over 2 300 participants right now um and it's been pretty interesting to see like where these um like there ends up being the strongest Community Support it's interestingly in places like turkey um Ukraine South Korea so places that some of them are very well known crypto hubs others less so um and there's been just organic traction happening from those places for the Atlantis program so that's a really easy way to get part involved um in the State Community uh so definitely would recommend checking that out as well fantastic all those links in the description for the listeners that's

easy access for you all but thanks a lot Jay this is this is this is awesome really appreciate you coming out awesome thanks Dan yeah thanks for having me on guys this is a ton of fun and anyone that has any questions can just feel free to DM me as well