Episode 350 - ZK in Solana with Mert Mumtaz

Transcript

[Music] welcome to zero knowledge I'm your host Anna Rose in this podcast we will be exploring the latest in zero knowledge research and the decentralized web as well as new paradigms that promise to change the way we interact and transact [Music] online this week drun and I chat with mumtaz mert is the co-founder and CEO at helus and someone who is known on Twitter as one of the most vocal supporters of the salana ecosystem in our conversation we cover Mert's connection to salana how he got his start living through the crash at the

end of 20122 salon's recent regrowth and discuss its positioning in the larger blockchain ecosystem today we then take some time to do a survey of ZK projects emerging in the salana ecosystem covering ways that ZK has been used for scaling privacy as well as in the Bas layer to enable new functionality this is something I've wanted to cover for a while now so really glad I got a chance to dive in we wrap up with a short chat about the past present and future of speculative assets the impact of meme coins and what could be next now obviously this is a little out of character for the show we almost never touch these topics but since we do in this episode I will make the note that this is in no way Financial advice

nothing we say here should influence any personal choices you make about investing obviously now before we kick off I do want to just let you know about the ZK Summit ZK Summit 13 is coming up it's happening on May 12th in Toronto this is the second time we bring the ZK Summit to North America the first being all the way back in 2019 when we did one in San Francisco if you've never been to a ZK Summit I definitely recommend checking it out this is the spot to find out about the latest research the newest applications to find find out who are the most important players in ZK today it's also a wonderful way to get to know through the larger ZK Community the application to attend is now open if

you'd like to apply to speak you can do so in the same form the speaker application deadline is March 15th generally I recommend applying early we do expect this event to sell out and tickets are limited speaker slots are even more limited so do get your application in early and we'll get in touch with you very shortly links as always are in the show notes and I hope to see you there now here is our episode with [Music] mt so today tun and I are here with mt mumz co-founder and CEO at helus he's also a salon researcher and a fellow podcaster welcome to the show Mt hello thanks for having me hey tun how you

doing good I feel like I'm outnumbered by Canadians for once true Canadians who don't live in Canada anymore yeah Canadians who don't let C that right nice so I've been wanting to do a survey about the ZK projects in the salana ecosystem for a while now I feel like most of the ZK action we cover on the show tends to be either in like academic ZK research it's on l1's dedicated to ZK or I would say like a large part of the ZK research out there has been in the ethereum ecosystem with rollups CK apps ID games all of this stuff but recently I started to hear more and more projects sort of popping up from

like in salana that are at least touching ZK and I'm kind of excited about that like I've been wanting to explore some some sort of like zkx salana for a while so yeah this is why I wanted to have you on and hopefully you can help us get a little insight into what's going on over there yeah I'll do my best not an expert but we do work on a lot of the teams that support it so yeah let's see what I got cool before we do that though let's get to know you a bit tell us a bit about yourself and I know you're one of the people who went through kind of the salana crash of 2022 so I'd love to hear a little bit about what kind of where you got your start and maybe what it was like to go through something like that sure before

that incident I was at coinbase and I believe I was one of the first two Canadian folks actually hired by coinbase so they had just set up their office and there I worked on a lot of stuff but mostly they have an internal system called Ledger which keeps track of all balance changes across all chains and transaction histories and all this and so I actually worked a lot on ethereum polygon uh some early optimism stuff and then even Avalanche salana was basically what I worked on on my own time cuz I was curious I was even curious about polka dot and a bunch of different weird things even briefly looking aland so I I was already

building on all these chains on my own time just basically trying to see what is possible with this Tech like I had been in crypto from a let's say investor point of view for a little bit but not necessarily Hands-On engineering wise as I was doing that I built a lot of stuff on salana and those generally tend to get the most traction these are all basic stuff like salana didn't really have any block explorers back in the day in absolutely zero data analytics sites so I would write my own indexers and then try to see if people were lying that were just posting on Twitter during the nft season you know they would say something like oh you know this mint is going super fast uh and then I would

write a script and be like well actually you have 99% of the supply and so it was it was very sketchy stuff and while doing that one I I noticed that the developers on Salon were super friendly and I just found most more of a community there than anywhere else personally the second point is that I noticed how difficult it was to build on salana so you know there was no concept of well I mean there is a concept of an ABI which is on evbm but on salana that wasn't standardized and also people were pretty Cutthroat and what that meant was to understand what the data looked like you'd have to reverse engineer it by very obscure means like Lally reverse engineering

bite code and brute forcing it or clicking somebody's like bundled JavaScript code in the front end unbundling it seeing how they uh set the types up and all this and uh I figured you know probably if I have this problem everybody's going to have this problem and so quit coinbase started helus raise around and then a few months after FDX happened oh wo so this is 2022 I'm now from that I can kind of date when you're playing with it is it like up like leading up to this so 2022 You're Building these block explorers and like hanging out in there and then at the I guess a few months after you create the company boom yep okay I love that idea though that you're like reverse

engineering like blockchains at that moment it's it's super frustrating I think for like the average Dev but I bet it's a lot of fun too because like by doing something like that you're getting to know the logic underlying like you're kind of getting to know the logic that the client devel Vel opers are thinking about and that they're kind of putting into it I guess yeah yeah it was I mean it's fun to like post about your findings uh because nobody else has that information so there's that inherent uh Joy let's say but then once you actually do it and then you know the client date uh client uh Dev just updates their contract or something uh or their logic which happens all the time your code is just totally broken and so you have to it's like a it's like

a war you have to keep updating your own parsers faster than they can uh update their own code which obviously is a losing game for me um but uh Salon has gotten much better at that over the years when did you actually first join like when you were at coinbase what what era is that uh like $8 Salon $8 um what year is everything is measured in price not time we're measuring price is the real measurement of time wow actually yeah that's a bad answer cuz salon's been $8 twice now um yeah uh okay so the if people remember the dgen Apes mint before which salana was at a very low price like below $8

and then this mint happened and then all of the nft season started happening that was around the time okay that's 2021 or January or something okay no no it's 2021 summer summer okay I see I see I wonder did you see salana early early on because I remember when it was first around 2020 2019 I think I remember like near and salana and Mina like there was a group of teams that were all sort of more Silicon Valley they were coming like just the engineers the types of Founders kind of who the investors were there was like this this generation um and salana was part of that I I wouldn't have been able to say back then this is going to be the one of

that Bunch that really like comes out flying do you think there was something unique at that early stage though that set it on the path or do you think there was sort of some like fortuitous accidents that led it to get to where it got I wasn't there at least in an involved capacity during the very early days so like before it was even called salana but I mean I guess I went through something similar when I was researching the different chains I want to build on and so I looked at you know Avalanche was pretty big at the time um near was kind of there um but they weren't as loud so me and my co-founders at helus we just we all did the same

program in University it's called Apple math it's it's a double in applied math and uh Communications signals engineering and so it's it's interesting because the way that salana is structured is kind of like a giant radio and so that just made the most sense like it just felt like the least amount of leaps that I had to make mentally to understand sand and I just knew somewhat scale I didn't know like there there's been a bunch of new introductions since then but I figured like okay for ethereum to scale which I understood relatively well at that time you need a bunch of conditionals in the future to evaluate to true and you're obviously they're obviously conditional on the other one so like uh you know the merge

had to happen you know 4844 had to happen uh uh you know Dan sharting has to happen you know rollups have to be interoperable all this stuff and while I think that will happen the simpler and easier bet for me at the time was like well salana scales it's broken in many ways due to how the code is written but that's a very generic problem right you need to make the code more efficient and and and you know add introduce some other things that we can talk about later but it just seemed like a less risky bet and a more interesting scaling path let's talk about the crash time so you just started helas a few months before that then the price drops did you ever have a moment where you were like okay I guess that's it no no I was it

was the total opposite for me and that's why I was so confused all the time on Twitter because I was already reply guying pretty aggressively by this point but yeah I was going to say we we got we we probably should talk about the history of how you became the the the reply guy of the era sure sure yeah so I was studying salana on my own time but I also understood how the e kind of evm chains worked at a high level and I was you know doing work with my own hands on both of them at coinbase but also on my own time but then I would always see I don't know when I started seeing this but actually I used to reply guys salana guys first because they would try to like say dumb around

you know the nft supplies like I mentioned and I would be like no that's incorrect and so I would it would started me as me reply guying influencers on salana and then at some point I noticed some people that weren't even really on ethereum but they were just weird fund manager types I think like Justin bonds is one that came up quite frequently at back in that time and he would say something like you know salana has a fraudulent design because you know the leader schedule ahead of time and so it's not random and I'd be like that is actually how a very large number of blockchains work this is not specific to salana eth 2.0 even is like pseudo random it's not even like purely random uh not even aware of anything that's purely random other than algo

Rand which has the name Rand in it and and so anyways I would just find all these posts and then and reply guy them pretty aggressively and then people just started tagging me and them before I was even aware them and so at that point once I see it I'm like a and I have to reply I know better these days but like back in the back in that time I didn't and then one particular uh criticism was always around FTX and Alama and these predatory tokenomics and whatnot at the time which brings me to this which is when I heard that FTX was going under I was like oh sweet like now the Predators are out now we can actually get to work and we don't have

this big piece of fud but I think I my time Horizon for thinking that was just too different to everybody else CU everybody else just figured no no no you're dead like you that's was the only reason you're alive which I didn't agree with and like people also weren't too close to the metal of what FTX was doing back then the last thing they released on salana was an nft Marketplace that totally tanked and and was just a total failure and they were also investing in like the move chains like sweet aptose they invested in polygon uh and a whole bunch of other chains and so it didn't seem like they were actually that useful for a very long time uh at least by the time I joined in uh you know mid 2021 and so I was happy we also obviously run

a company and our numbers were not affected at all okay we were still getting a steady stream of developers using the product uh and in good growth week over week month over month and so I was just very confused and at that point the amount of fud on Twitter got to such a level where it became basically my second job to just reply to people and be like this is this is incorrect you are extremely incorrect or some combination of those along second second job or first job arguably there was a there were times it could have been the first job and people ask me this all the time like what do you what is the split between Twitter and and and um being the CEO and it's like I see it as kind of the same thing like I don't consciously

schedule like okay I'm going to do calls or I'm going to do podcasts or flights and conferences it just all goes along with the same job because if the ecosystem is dead because nobody's actually refuting these points then you're not going to have a company to begin with in that ecosystem at that time everyone I mean so the fud was crazy the price dropped I what I remember is that like all the prices dropped everyone everything was affected but salana dropped a lot it was kind of I mean I think at the time did it get down to eight actually in that little block down from something like 200 almost or 2 yeah wild yeah like there I I'm guessing there was sort of the Diehard crew that stayed that continued

building you were talking about like you still had the same number of customer customers who were those people at that time like how would you describe that group backpack is one so uh with armman and and Tristan at the time they had just released a new crypto wallet uh backpack and then this nft projects madlads for example that was one I mean Jupiter was still around all the dexes like well not all of them but Orca radium were around Camino was around jto was around I mean actually not many people left I think some people tried or experimented cuz I think Aptos launched kind of right after this but like that didn't work for them and so they came back from Aptos there was also a lot of

drama with nfts and uh them getting grants to go to polygon and at the time I had looked into this for a little bit and I was like wait a minute these guys are just getting paid to go on this and so I called that out and that got me into a whole lot of trouble for for a very long time and then I became actually very friends with the polygon guys okay uh but so like there was a lot of those Shenanigans going on but really like with the exception of Doods which was just an FD project nobody really left they all kind of just had worse revenues and activity but I didn't see many people leave the main Builders stayed tun did you work on salana stuff at that time uh mainly the main thing I had done

at that point was kind of I invested in J and 2022 and so I'd spent a lot of time just like reading the scheduler code and trying to understand meev stuff because one thing I think I think the thing that convinced me there's something I Mis assessed uh salana over time was trying to really understand the scheduler and the me usage and uh how kind of like they it has a very different mechanism for priority auctions and priority fees and it was sort of a a sign that people were really using the chain enough that you had to come up with some of these things whereas like you know you look at Cosmos chains or you look at Polkka dot

or things like that you could tell there was like no real usage because no one had to build any of these priority scheduling things there was like not really like this type of like hey we have to optimize our meev auction for our chain because we just didn't have any users to do that there were a couple me projects I know on Cosmos but they never really took off there wasn't enough or they pivoted or they pivoted yeah exactly exact but my point is like there there sort of um I'm trying to remember the name of this like kind of apocryphal law it's basically a thing that says like the structure of your code base reflects the structure of your organization like if your code base is very hierarchical it means that you have a very hierarchical organization if your Cod is very random it's like oh you've

just had like people come and go quickly and there's a sense in which like when you look at how people think about the transaction supply chain from like end users all the way to the chain that tells you something about how much people actually are using the chain more than the actual code itself like if they've had to modify these types of things it's because like they've had kind of crazy demand and usage so I think I at that time I was just trying to understand this meev related stuff um because like yeah the auction model in salana is very different there's not like this one big auction like there is in ethereum there's kind of many small ones that get aggregated and I think in general the the kind of high throughput nature of

salana just like makes it look very different yeah I I think like it's good to have had like to have both kind of the like horizontal scaling versus vertical scaling blockchains cuz like in some sense I think there's probably not going to be like a winner takes all that wins in in one of those but that may maybe Mt will disagree with me on that but no no that sounds right one of the Mis uh big misconceptions that I saw quite a bit on Twitter was like a lot of hate towards salana because salana claims that it can do all the traffic on the universe or something around crypto but that's actually not the case and that's not the goal uh the goal is just we think it's cool to have a giant State machine that can sync information as fast as possible around the world that's

it now that'll benefit some use cases and it will be totally senseless for other use cases so uh it's really about what are the business requirements here why do you actually need a chain and actually just bringing actually we're sort of we're still in late 2022 early 2023 but like it's evolved a lot since then has the tech changed do you feel like you're dealing with a different Beast under the hood now I mean obviously tons of tooling I think you've built some of it so yeah like tell us a little bit about that Evolution yeah well people probably remember all the outages in 2022 oh yeah and there wasn't many in 2023 it's actually uh due to what tun was saying there was a lot of demand so

like there would be an nft mint and people and I mean these people were insane they would get data center grade equipment and just Spam the living out of the nodes uh like 10 100 gigabyt per second bandwidth times few multiples of that constantly and so the amount of traffic that these nodes were handing was pretty insane and and so we knew that that wasn't going to work like toally and and team basically did not expect how generate people would be would be uh when it came to actually using blockchains they thought these people you know must be somewhat reasonable uh and you know they'll just do you know order book stuff but then he's like you know why are people just

uh you know aping this much into like jpegs and and so there was that kind of whole series of outages and there was a few big changes that came from that that most people just I guess weren't aware of but like one was the introduction of like changing the networking like going from just raw UDP to quick and like Google's thing uh adding um stake weighted qos quality of service based on stake bandwidth or getting bandwidth in proportion to stake localized fee markets right people don't actually know this but salana has said they have fee markets but they haven't actually had fee markets in production until like maybe not even last year so they're actually quite new and and a

bunch of other changes like that uh uh some of them quite boring but very much Dynamic so for example just this past upgrade 2.1 uh which is the latest client upgrade they made a a few changes to the client such that because you know we say the block times are 400 milliseconds but in practice they can be like 450 500 milliseconds even and now they upgraded the or or fix so many bugs in the core client that now the average block time is like 390 milliseconds so it actually went down and that just goes to show you like how much it contrasts to ethereum where they did not plan that to happen they just wrote the code fixed it and then the blog time just went down

uh which sounds good maybe not so good who knows uh but there's a lot of upgrades like that and then going forward there's a much more structured process because now there's a second client with fire dancer and a much more structured road map versus the early days you can maybe think of it as like a startup just not trying to die yeah so so actually I'm a kind of question that has come up in and caused fights in ethereum but I've actually been kind of curious about how it is kind of viewed in Sala is you know the multic client road map in ethereum has been controversial for a bunch of reasons one you know you have five clients if you have a bug in one then like how do you replicate the bug in the others right like they're very different code bases

different languages you might not be able to perfectly replicate or be bitwise compatible and we've seen a couple bugs live in in ethereum where that has been true isn't it supposed to be that if there's a bug in one it's saved by the fact that there is no bug in the other that's always the story they tell that's one side of the argument but the developer side of the argument is I have to maintain multiple clients right yeah so you would would you try to introduce the bug if it's not a breaking bug just to keep it going well I I I I I'm more just saying like hey like maintaining five copies of something that's supposed to be equal to one has a cost and the development cost of like I add a new feature to a fork and now all the clients have to agree

there's some overhead right so like for instance in the fire dancer and Anza Agave type of thing if I make a fork I need to make sure my the forks agree beforehand and sure I could go run them live and run into the edge case but that's not a ideal scenario for a so like given that you are running infrastructure provider you're running rpcs and making sure people can how do you feel about this kind of like multic client thing like two is you know not 10 but like yeah where where do you what are your thoughts on the multi client for single client tradeoff uh yeah I mean just from a purely selfish point of view we don't like it because it's a lot of extra work for us and it even happens today like with the orderflow games

because the protocol isn't necessarily defined very well at the boundaries and and so there's like six different ways to send transactions today where the RPC somehow has to figure this out now the the reason why fire dancer was extremely important especially at the time when it was announced is because of the reasons I said which is the software kind of just sucks on the core client like it's just an unsalvageable code base it's totally broken in many ways and the protocol has no spec like a lot of missing basic things that are extremely difficult to do a refactor of but are probably doable when you do a rewrite in something uh as a fresh rewrite and so that's why our fire answer was exciting

because now with a pure rewrite and now good uh interface boundaries and specs and all this that's definitely a net better uh situation than the existing stuff but I'm not sure what happened here but Anza hired a lot of very smart people to the point where those guys actually it seems like they've actually been able to salvage the code base and it seems like they probably will be able to make it much better going forward and so now you have this case where you have two clients who are probably pretty equal I would say especially because fire answer hasn't been live yet and so I'm going to give the point to Anza in that point uh in that case and so now you have the a much different scenario than ethereum where ethereum is a bit

more like you know uh equality of outcome in a sense like I don't I'm not going to get political but like it has to be that way for the system to work where on salana it's like they're kind of just competing against each other as hard as they can and whoever gets the most validators who adopt it is kind of the the winner and that also depends on like how well are they serving their customers who are the validators right like are you maximizing uh the the revenue are you are you actually making transaction inclusion much better less fragmented all these different things and so I don't know how it plays out we certainly don't like or I don't know many people on salant who actually like the concept of multiple clients past maybe two and the only reason two is like useful is because like a day Zero or something in

in one of the clients and then you're totally and you have to use a different client or something but as an RPC provider not a fan and we'll I'll let them battle it out yeah I mean I I only ask this CU like in ethereum I actually think the interesting thing is that the production clients and the testing clients are completely different like a lot of people like developing with W but they don't they use you know no one who's running execution nodes is using it and so I I feel like there's sort of this problem of like too many clients equals too much maintenance obviously yeah from a from your perspective I totally get it more than two sounds like a nightmare right but I yeah I just there's sort of this weird thing I've never really understood why you need so many clients I I'm I'm I'm

more of the mindset that like you're more likely to have developers introduce bugs by accident when they have to rewrite the same thing five times in five different places than uh otherwise but interesting I actually hadn't thought of that what you just shared M that observation of like the multi client and and roughly evenly used multi client like that's kind of that's seemingly a bit of the goal you don't want one to win in ethereum and you're kind of saying actually in salano would be fine if one one yeah I mean like and there's some architectural decisions or protocol decisions that affect this I think like ethereum has the concept of an inactivity leak for example that might play nicely with multiple clients whereas on salana yeah I mean we don't

like outages nobody does they're unacceptable but like if there is one like how much do we care about safety overi or what's that tradeoff look like and obviously everybody has different opinions on this but if there's one client live and then you're doing it such that you have a backup running with the other validator so like helius were the biggest validator assume you're on Anza and then we have in the background and like a blue green deployment fire dancer running then let's say the chain goes down from the Anza bug we can switch over now it might not be instant it might take 30 minutes or something but that still just did the job I thought I heard and this is because I'm not that deep in this ecosystem but

didn't when like the Trump token launch wasn't there some sort of like outage or problem or something didn't quite happen like maybe it was only on the wallet level but can you share a little bit of what happened there sure so interestingly the Trump coin itself uh was handled very well by everybody what the problem was uh within two days Milan launched a coin and uh when Melania launched a coin the following happened this is kind of hard to follow but roughly speaking there's two paths for transactions on salana one is the regular salana clients one is jto right so jto and salana let's say jto is more used for you know obviously Meb folks but like Arbitrage Traders and all

this and what happened was jeto became degraded and when jeto becomes degraded those Arbitrage folks now can't use jeto and now they must use the actual main schedule of the chain which means now there's much more competition and congestion on the main chain which means among other things the fees go up but also since the schedule itself is so inefficient it inherently makes the chain less performant like you process less transactions but also uh some very bespoke things here like you know uh Jupiter for example hard caps their fees on the front end such that as a user assume now the fees on salana have risen from 0.001 to $1 Jupiter will not set

the dollars uh the fees to a dollar they just won't do it there's a hard cap for safety purposes but but if you do that then your users are literally locking themselves out because they're not paying the fees they can't and that actually tends to be the case with all apps on salana including Phantom and so you have this situation where everybody is bidding below market for getting included except the Arbitrage traders who are Bots are they blocking them out are they like taking that away from the chance that like the users can't actually get anything through is that was that what what had actually happened that like people were just trying to get a transaction and it just wasn't happening yes and but but the reason is yeah I mean one is

obviously jto went down which caused the fees to be higher and caus more contention on the main validators but the second and I confirm this myself is the apps themselves not intentionally a force but accidentally locked the users out from competing because if it costs $5 to be included but you don't let your users pay over 5 cents no you's going to get in it does and so what I did was I went on Jupiter mayly change my own fees and then bit above the market and I would always get included myself but like you're not going to tell all you know thousands of users hey you have to set the fees manually now that's just not going to work especially not all the new users that had just joined for the first time yeah yeah so that's what

happened but like uh there was no uh halto block production or anything you could I think probably make an interesting argument like hey does jeto being degraded mean salana is down because it is a jeto client but I don't think that would compute if you took it to its eventuality is there actually like a postmortem on jeto though like is there going to be a community effort to fix that or does it seem like such an unusual Spike and they don't expect it like anytime soon so it's okay to leave it as is no no I mean those guys ship I mean they haven't disclosed everything but they have many upgrades coming got it all right I want to dive in now to some ZK stuff um I feel like talking through the history of your involvement

in salana but also just Solana's development this has been really helpful also to share a little bit of the distinction between salana and some of the ecosystems that we're used to exploring ZK in let's do this survey that I wanted to do which is like what is out there in ZK and maybe let's start with like what was first where did you first see like ZK being used at what level is it is it like in the kind of roll up I know it's not a roll up in salana but is it like roll up like is it for compression let's say or is it for something else like privacy yeah I think the first and I could be wrong I'm sure there's like some other thing that I haven't heard of but if I'm to just maybe put on my mainstream hat probably the first two

were um Elusive and light protocol and those were both doing privacy at the time okay so private payments private dexes things of this nature and there was even like an nft project that was relatively popular at that time I think it was called like dark protocol or or something along those lines and they had you know some small ecosystem of tools like ZK confidential transfers and stuff like this and so those were kind of the big ones and then they basically all pivoted uh because I don't think they found product Market fit with privacy um and so elusive rebranded to something called

archium and uh now they do what's called like verifiable compute um where it's kind of like a different network but it still uses salana for some elements and then light they pivoted to basically ZK scaling but this looks very different than other chains because Salon scaling these are very different so to give you a rough idea evm chains generally speaking not always compute is relatively expensive but storage is cheap and on salana it's total opposite where computer is extremely cheap but storage is very expensive and so for example if I send you right now an nft or usdc what I need to do is I need to pay to open your storage account for

that asset so not only do I have pay a a transaction fee for for doing computes I also have to pay a fee to open up your storage uh and you can imagine that let's say you're doing an airdrop I now have to do this to 10,000 different accounts and so that's 10,000 storage fees and then 10,000 compute fees and so we had this thing called NFC compression uh this is before light where basically what we would do is we would not store the data in these storage accounts so we the data would not be stored in accounts it would be stored as let's say call data on transactions so instead of opening

10,000 different accounts and air dropping to 10,000 different uh people what you would do is you would do 10,000 different transactions and then the details of that transaction were in the form of Mer leaves and so then what you would need is an infrastructure provider like helius to replay that state and then expose the latest state to people okay I like the distinction you're making though the compute being cheap on salana and expensive on ethereum I mean I'm obviously way more used to the ethereum model so but I don't know if I would have known that that that you'd consider compute more expensive on ethereum and storage being cheap I mean I feel like what you've described here it suggests

that like rollups how ZK rollups are working are they making use of that cheap storage on sort of the L1 and so if you try to replicate that you'd have to kind of think about it differently if you're doing it Sal well so if I'm to maybe extend the thought from before with the compression stuff so now what happened is the compression stuff that I just mentioned and this didn't include ZK only worked for nfts it was specific to nft accounts okay right because on on ethereum your your your uh data and your logic is in the same contract it's in the same place on SL that's not the case you have explicitly separated out of accounts for data and logic

right so it's decoupled in that way and so then the question is like okay if this works for nfts can it not work for all types of accounts on salana so generalizing compression to the entire chain such that we scale storage for the chain right we're not scaling compute we're not bottlenecked by TPS we're bottlenecked by rent which is the fee that you pay to open an account and so with ZK compression it is any account on salana you can compress it and then link it back with a miracle hash or a miracle root and so you can guarantee the Integrity of of that data on The Ledger but so this is this is kind of funny so

it's not a rollup then cuz like I for some reason I thought I thought it was like an a salana equivalent to a roll up different but similar in that it's like batching a lot of different state but here you're saying like it's a single account being compressed or many accounts but yeah it it is not a rollup and and this is this is where it gets really interesting or funny you know at the time we're like okay what should we call this and um the ZK is not used in many places here ZK is used only for the proof and making it constant size like you can actually get away without using ZK uh so it's kind of like a nice to have but it's not strictly necessary but but we call the ZK compression because

we use compression and then we use ZK on top of the compression so we released this product by the way helius did not build this we only Built the uh indexers light protocol built the actual protocol okay but so I tweeted it and my goodness everybody on ethereum was like oh my God they're just calling it uh it's a roll up it's a it's a validium it's an OP it's all this stuff and we're like no no no it's it's not any of those and nobody would believe me and so what I had to do uh was I was on farcaster I was like I need to get a hold of falic somehow and so I described how it works to vitalic cuz by the way we always call this the technical term for this is a stateless stateless accumulator right this is not a New

Concept this already exists in computer science it's a stateless accumulator but on ethereum people kind of conflated with the rollup and so I'm like metallic what would you call this and he said yeah this does not seem like a rollup it seems much more like ethereum's like you know verical tries and like stateless client architecture stuff and I was like yes thank you so there's a lot of confusion there but it's not a roll up in the traditional sense of what you would call a rollup got it yeah I mean I think the controversy here was kind of funny because I feel like there are other things that people call Network extensions in salana that are much closer to rollups than this and and or are just basically rollups um I I don't exactly get why

everyone wants to fight about the nomenclature though cuz it's like don't you care about what it's used for rather than like what word I call it um and somehow I feel like this entire discussion made me feel like remind myself of that cuz it's like is the name that important I think isn't it what what it's used for more important but that was just sort of my feeling when I was reading this and I muted the word rollup for a few weeks I feel like um so but you are saying there is something there are kind of like equivalent to rollups already in a different form but maybe not using ZK yep yeah are rollups on salana I don't think any of them are alive okay um there is so maybe going along with the survey from the past

question so like succinct is now on salana uh RIS zero is on salana they have you know boundless and they're doing some other work there's bonso which is by anagram that's another ZK framework and then with within salana actually has no live rollups today but one team that's working on it is Zeta markets their new products it's called bullet and it's a sovereign rollup using uh Sovereign labs's framework so that'll be like the first Sovereign rollup on salana so the term Network Extensions by the way horrible name makes no sense and I tried telling this to the folks who came up with it long before it came out and and the reason is simple uh a name

is supposed to be specific such that the reader understands what's being referred to but uh in this case it's more used as a marketing term to mean we can come up with with their own names as well uh and so the easiest way to think about it is uh a network extension can include al2s but it doesn't have to it can be many other different things it can be for example ZK compression which is not a rollup yeah but it's a network extension okay that is also I see so it's like the meta category then yes but doesn't the whole idea of network extensions kind of go against what salana build itself as in the first place like going back to that story of like how ethereum you know the the full extent of e theorum is the merge the eips the rollups the talk to

each other and having those extensions off the L1 it's built because of the way the L1 works but salana I feel like was always build as like you don't need rollups because it's so performant as an L1 so like why have rollups like what's the point there or sorry Network extensions or any of these things yeah so the the difference here and toly will make this distinction all the time is that what we care about is having the execution on chain right with with a traditional rollup on evbm for example there's not enough compute that you can do on the L1 so you must do the execution offchain and offchain me means a roll up and then you roll it back up to the L1 right so execution is not on the L1 but we think that execution is the most interesting

part of salana or let's say blockchains and so as long as execution remains on the L1 which in the case of like ZK compression for example all the transactions are still on the L1 so the execution is still there but the latest State exposed might not be on the L1 but that doesn't matter because sana's thesis is like the value is in execution right me Priority fees auctions all this requires execution and that's what we want on the L1 now that being said there's another like slippery line of thinking that I've seen recently which is like okay well salana is getting some Network

extensions does that mean it's becoming like ethereum and I would actually say it's becoming more like Bitcoin where the the core difference here is that the L1 does not adjust itself to cater to rollups right like on bitcoin you can build a rollup and and Bitcoin people don't care you can Fork you know and put a dogcoin on it or something Bitcoin people don't care they just have that one thing they want to do salana is the same which is they just want to you know reduce latency increase bandwidth meaning keep execution on chain now if somebody wants to come and say hey I'm going to actually also build a rooll up on top of this they can do that because it's the permissionless network but like you're not going to get the core devs to

be like okay like let me now give you some extra blob space or something that's kind of the core difference what about then ZK VMS though because in the case of ZK VMS at least again the way I've understood it a lot of it is about like you're you're submitting proofs but a lot of the execution is actually happening with wherever but often not on the L1 is that different when you look at like succinct or risk zero like connecting to the salana ecosystem or is it are they still using Salon in the same way again it's a permissionless network and so if people think there's like some weird Market Gap where people need much more compute than salana and they think they don't want to do it on salana and they want to do it on rollup

then yeah I mean it's a free market you can build this product see if you get traction all this the core difference is again the core devs are not going to hear um they're just going to like they're not going to write new OP codes or something just so that you uh ACC company rollups like for example the big difference between e ethereum is that if you look at the road map it's designed to become a hub for rollups yeah right salana would not do that they just care about the one thing which is ibrl and uh now if the block space is valuable enough such that you want to extend it and and build a roll up on top of it by all means go ahead and and you know like for some cases it might it might be better for that like a base rollup right

with much shorter block times that might be better but but the core thing is that like jeto helus you know fire dancer Anza we're not going to build rollups like the cord devs just won't do that I don't know if you followed any of the ZK development in ethereum but there was also like the co-processor idea this idea of doing or like you know sometimes it'll be like an AI environment that lives offchain and then uses ZK to come back but does that actually matter in a in something like salana because like yes it's amazingly performant but I'm guessing it's still like blockchain performant is it able to do really big things really big computation loads is it able to do like something that would be yeah you'd have to have like in the

real world like tons of machines running it like can it do it no um I mean obviously it would depend on what we're referring to but I would say performing in the context of blockchains not performing in the context of compute in general yeah it always comes back so with salana it's very simple because there's really just one Vision uh which is decentralized NASDAQ right can you sync prices across the world with minimal latency right that's it now if you want to do other things on top of this then you can so with uh with you know RIS or you can do this actually son has the concept of uh ephemeral rollups where state is extended ephemerally and then brought back right after the yeah and so there's like funky things like this cool and like I'm being a little

repetitive but the whole dichotomy is that people can do those things and salana like you can't I'm not going to be able to ban anybody hey like you know you're breaking the law here the only thing is we're not going to change core yeah to accompany it are you sure it'll never change how come you're so sure for for sure it could change in the future like I it wouldn't cuz otherwise it's just an ideology and that's not engineering right uh and so engineering is just just about okay what problems are we solving that actually exist and that we actually can solve better than others so I I I don't foresee it changing and at least the next five six seven eight years maybe because you can

also see the uh pent road map which involves like async execution multiple concurrent leaders and and and other things along those lines that being said you know let's say some breakthrough happens that we don't see coming and within two years now you must have ZK on chain oncore to an enshrined concept or something then you know we're still pragmatic we won't just say no no no like we're you know the gods told us not to do this so we won't do it uh yeah well I I another thing I will say is like I actually think the co-processor type of stuff makes more sense in some ways and I think that's why the ZK you see at least from people

taking the existing VMS and applying it to Sal look more like co-processors than it is is like a roll up in the sense that under really high latency constraints if you want to do some sort of like long running calculation infrequently you kind of don't really want to constantly be broadcasting that to chain and dealing with kind of getting the transaction in but you do maybe want to do that infrequently and in theory these Salon of validators are already running beef Hardware so like maybe they do something like reaking like or or something where they like are participating in these offchain co-processing networks alongside their validator and only infrequently have to post so I think in ethereum the

co-processor stuff never took off because there just wasn't enough demand for it and people once they went to rollup they just kind of were like okay well a compute so cheap I don't care anymore it was like sort of a very weird thing uh so I think that's why we also saw you know in the market every single co-processor just pivoted to becoming a VM which was also just like a weird like the idea of like I had a product and I'm going to become less of a product I'm going to become a virtual machine is like a very funny thing to observe about a market right cuz it started narrower and got broader usually in startups it's like I start broader and I got narrower well they moved to a different part of the stack almost they were like oh okay we haven't quite nailed it but actually

what did like garage moved a different way they went pro network from co-processor I think that seems to be the two directions I forgot one thing uh about zon salana which is that because I was talking about how Cavs want to company it that's not technically correct in in the full sense because salana has uh these new things called token extensions it's pluggable uh programs into salana and one of them is a confidential transfer So within salana core you can actually do confidential transfers and those confidential transfers use ZK so it's built into the core protocol that you can confidentially transact I was also reading about something maybe this is

what you're talking about but it was called like CIS calls and that they're were some like ZK related ones is that what you're describing or is that something else it's related that goes like pretty in the weeds but that's not like in the context of scaling it's to enable other things um like confidential transfers like but this is inside the protocol itself so this is where we're like I'm guessing yeah that's cool can you can you explain what those are in my for our audience like I think they'll they'll probably follow uh assuming the audience is like evm focused a good way to explain it would be on evm you generally need to deploy a new contract for a new token right and then you need to write different things for that token and then it needs to follow the erc20 interface on salana this is not the case

on salana you have enshrined programs so like the token program and then you just say token program create new token here's a ticker so that's like very simple very different and so one of those new features is that token program create new token and then transfer confidentially so now every token every token with a token extension on salana can be transferred confidentially so you can see who received it and who sent it but you can't see the amount that's so cool that's like partial privacy all of a sudden but built into like when you're this is like is it a choice when you're actually developing the token like will the will it be individual tokens have that quality or is it like certain account will have that quality yeah yeah

you can uh you can add this to any individual token oh cool wow yeah I haven't heard of something like that and that's ZK yep nice I feel like we've done kind of an interesting survey of like where ZK is being used in this stack some of it is applications on top some of it is actually within the the client itself but you we know we started this talking a little bit more about like the m in salana and like the fact that that was really active compared to other ecosystems I know that in me land tun you're often telling me like tees are used for me stuff there's reason to use other kinds of either privacy preserving or Advanced cryptography do you know Mt if there's also stuff like that happening in the salana world where

yeah trusted execution environments are used in order to I don't know hide something to make sure that there's this place where execution can happen without anyone seen it on chain uh I cannot comment on this I think per perhaps uh in the future okay is it because maybe there's something but you don't have to say sounds like something might be in the work once it's released I'll I'll flag it to you so you can I I I will I'm in the same boat as Mt actually oh like sworn to secrecy that's amazing let's just say PE people are in Sal are looking at everything it's not like they're like you know hey we hate cryptography or something can I I mean there are these

other you know sort of advanced programmable crypto stuff like MPC and FHA both like FHA especially is still very fresh so I don't know do you have you heard of anything like that being used in the salon ecosystem MPC FHA I mean even I think you did mention one thing the the hosted MPC wallets like the turn keys of the world are certainly used in Zana right like if I look at moonshot Moon shot's basically enabled by NPC like login um so I think but that's not something that's unique to salana right that's like all networks kind of have that yeah yeah there's a few mCP teams who use mCP like MPC MPC um no no no no but the reason he's

saying mCP is that's multiple concurrent proposer which is like something that's getting implemented in yeah um it's actually embarrassing because I worked on NPC at uh coinbase uh one coinbase wallet products but yeah on salana there's a few teams there's like squads TurnKey different wallet abstraction folks backpack is adding it for their exchange products which is onm uh Cube I think uh actually is also like a hybrid L1 Dex sex that uses MPC pretty strongly as well so a lot of usage for MPC specifically for FH I'm not sure uh I think Kyle is pretty interested in

that like multi point but I'm not sure if I've seen anything I'm actually going to link to a Blog post from helas which is focused on ZK applications in there you did mention that archium which was previously elusive was using something called multi-party computation execution environments which sounds like a combination of mpct but maybe ZK too um but yeah do you think is that related actually do you know I think they basically rebranded that to verifiable compute so uh I don't know exactly what they're going to do like they go toet Market isn't um I haven't kept up with it but that's that would be a team that strongly uses ZK and other cryptography on Sal I have a a

more practical question which is you know we've talked about a lot of these applications but I'm kind of curious what what is the developer since you see developers who are building kind of you know using you to submit transactions and test and stuff like that like what are the kind of features of the vein that use Advanced Cy that you see demand for like are there a lot of people trying to do private tokens like private transactions or is it more focused on compression and saving gas like and doing kind of these like token launches becoming cheaper like what what does sort of like the demand on the developer side look like and that's today because you did say that earlier there had been no product Market fit for privacy but yeah curious if that's changed yeah I

would say there's probably still no real pmf for privacy on salana at least although I invest in privacy stuff pretty frequently I the majority is based around compression which is just scaling uh large scale bulk token operations or account operations and then yeah we've worked with obviously light and yeah like there's CIS calls in the protocol itself so like the developer experience is not too bad just directly working with core and then Sovereign Labs that's been an interesting one although I'm not sure if they use ZKA for their stuff or if it's just for posting da um I'll to double check that and then like there's some interest around you know the scaling stuff that I've seen but generally we

don't work too closely with those folks cuz like uh or or at least I should say that wouldn't technically be salana this might be for their own L2 or something which is an svm rollup but um that doesn't sell on salana but I haven't seen many selling on salana cool well thanks for that survey that was actually really helpful ZK in salana does seem a little bit at an earlier stage like it just it sounds like there's like little inroads but you just kind of mentioned maybe it's not being used that much yet like I don't know if the utility is completely found within the the ecosystem yet yeah I would just say that the teams who actually work on those like light would probably put me to shame on that like they focus quite a

bit on it but from you know our customers are regular app developers mostly or quants and Traders and stuff so is entirely possible that we just don't come across those types of users and and so light would have a better idea okay that being said um Phantom I believe is adding uh ZK compression tokens today so that goes to show you like it does have adoption um in like certain projects were using it for their tokens as well nice now I just have a question just generally I'm just curious and really it might be the corner of X that I'm paying attention to but like meme coins and salana like it's a it's a narrative that has like moved together and I just

wonder like at this moment snapshot what's your feeling about it is it forever do you want it to be forever is it dead is love for is love forever this is like asking that the same kind of philosophical question I could ask this actually relatively frequently So my answer is pretty uh generic at this point and so okay basically if you look at the history of crypto the thing that has the most pmf is generally speculation in different forms right so you have I nfts meme coins dog coins L1 Forks all these different things I would even put governance coins in in some regard to that category yeah and so it's almost

always the same thing and you know we can debate for why that is and I think it's probably due to poor regulation but yes um that is the case like that is the big thing and so that's very general to all of crypto it's not so like when nfts happened for example they were on uh openc on eel1 and then they were on Magik on salal one and then they were on bitcoin and so they were everywhere with with meme coins um it's just a latest form of this and it it ended up being on salana because well I think a few reasons one is that I think creating tokens on salana is again you don't have to write a contract you just click the button that's it yeah it's easier Phantom is already integrated and you

don't have the temporary problem at least of this fragmentation between the different systems of rollups you just do it on salana and um so you take all those and then you know there's Jupiter um there's all these TG Bots there's pumped out fun obviously from from Sal's point of view it's like decentralized NASDAQ and what that means is we want to enable the trading of assets at the fastest speed as possible as affordably as possible in a decentralized system now the names of those assets can be nfts or they can be us DC or it can be memec coins and today it just happens to be the case that most of crypto is in the memec coin phase and most of that

happens on salana yeah but from our perspective as like core devs let's say we don't shair it's it's a really good stress test like we wouldn't have been able to handle the meme coin craze that just happened were it not for nft craze a few years ago that made us improve the core system to not go down again in in cases like this right there are some promising things things like the stable coin Supply on salana more than doubled since the Trump launch for example and then many more rwa start launching like tokenized gold Franklin Templeton just expanded slant today for example and by the way like other ecosyst like appal had a mem coin fund base is obviously very involved in trying to get that

going I still view AI agents as mem coins I I I said this on podcast the other day which is if you're if you're targeting anyone 25 and older you can go to l2s if you're targeting people 25 and under for meme coins you have to go to salana there's like it's a very age bifurcated Universe I hate to say I mean that's why I think all the L2 meme coins are things that are only funny to Millennials it's sad but it's true it's like it's just it it's just like oh yeah the age grab is real it's funny when you talk about it being the NASDAQ I start to be like so are the meme coins kind of like penny stocks like in the future is that what they kind of sure yeah although I don't know are

penny stocks on the NASDAQ actually are they like officially listed well there's like the OTC pink sheets whatever stuff okay yeah yeah I wonder like if there will be because that's another thing you just describe like how easy it is to launch a token I feel like in the past you could launch a token on ethereum pretty easy I mean yes you had to deploy a contract but there were websites that made it like almost a button but it wouldn't get listed it was impossible to find well I think the most important thing about pump. fund style mechanisms versus like all the historical stuff is like a it took a bonding curve so you you know you put in some money and then you get some fraction of um sort of the tokens and then once enough people put in money into bonding curve it converts

to an amm and then it burns the token so like there's no wallet owner of the lp pool so that means that like the person who created the token can't just remove all the liquidity and like immediately rug you now there are other ways they can do that but you know via sniping the pool and whatever but like just at least at face value it's actually a lot better cuz like there have been all these like social token things for years and whatever and they all just use bonding curves but the bonding curves would basically get manipulated and there wasn't like a free market to trade it so like the idea that the bonding curve converts to an amm pool was like the main innovation in my mind although of course like if I'm running me bot as soon as it converts and I snipe all the

tokens and I've basically got rid of that but it took people a while to you know a few months before I think that kind of flipped from like the fair launch everything in crypto you know to to mer point is like all of these things have something similar they all start with this ethos of like Fair launching right that like the supply will be distributed fairly and within a couple months the market figures out a way to make them not fair launches game it that's but that's like sort of the cat and mouse game of of of all tokens I think think meme coins have been around a lot for a long time nfts were around for a very long time before they became popular what do you think exists today that could be the next Generation kind

of speculative asset or speculative object mechanism I wonder if you'd spot it on salana could be AI maybe no no yeah people used to sit telling me Ai and one of my hot predictions for 2025 was that AI will die by the or AI tokens will die by and it seems like that's actually already happened uh so it took much much much I know tun loves the D stuff uh oh D will be itch D will true scams true scams please do not could be rwa they never really had their moment CZ is really Shilling D though so like I feel like I you know I'm probably going to get overrun by CZ oh man unfortunately I think it'll be um if it

does happen it'll probably be more social fire ented because I think this has been tried many times and hasn't worked yeah and so it's like probably somebody's learned from this mistake at some point in like how to do social fi less incorrectly and are you when you say that do you mean like the friend. tech kind of social fight or do you mean like launching a token into a social network because I also think that has never been successfully done like telegram has tried a couple times and like it hasn't quite worked but if like X was to do it maybe that would work well something that could be successful that's already doing quite well is like uh KO kaido uh with the yaps right that's like social fi in a sense you

talk socially and then you earn some points and those points presumably give you some tokens or something maybe so like social intelligence and like financializing that somewhat and I I know of at least like four or five big social F apps coming to Sal uh I think that clout one just launched um and a bunch of other there's like time fun so I think those have a good shot but I have no idea honestly okay I have I I have I have a question for you for a 2025 prediction which is how long do you think thef fun domain lasts does it still grow to the end of the year or does it fizzle off the way do XYZ kind of fizzled off that's a good question I think something that will make it keep that that will keep it alive is Solana

mobile launching and I think they will they will try to like be like oh you know we're very mobile focused and stuff and I think there'll be a new wave of those with the mobile stuff and um I think it last 2025 I don't think it last much longer than that but it definitely lasts a year in my view I also don't think it's full mainstream yet I think it has space there's too many things that have not been dot fun I mean social F does remind me of nfts in one way when I think about nfts in 2017 like crypto kitties like they had this huge Boom the way fch did and then it crashed totally the same way and it stay dead for like three years more than that it's 2021

really really like four years so dead social f i i do I kind of agree with you it could have a rebirth but it and people hate that though yeah maybe maybe we hated nfts too at the beginning you know one uh one thing that I think is super under the radar and has not really worked but I think could is like these sport tokens um like karate combat um there was that basketball one that like the crypto Kitty people had done what was it that was the initial initial nft thing in 2021 what was this NBA Hot Shots or something like this Top Shot Top Shot exactly yeah I think some of these Sports ones I know a few teams launching

something around this soon and it seems a little because the athletes tend to be very uh crypto forward let's say and um and I I haven't seen anybody really talk about it and so it's pretty under the radar and so that alone could be like oh wow let's not miss out on this thing that we just found out about so perhaps that's another option cool feel like we we veered off our serious ZK conversation into like what's cool next we we have we have we have we have the president of helium here as M would like to say yes people actually believe that that um the other day they launched a free phone plan and I put my phone code uh cuz they

gave me an invite code for signing up free and like serious you like VC started reaching out to me like hey I'm sorry we couldn't work together on helus but like helium looks pretty cool and I was like what the hell nice Mt thanks so much for coming on the show for sharing the story of you know your involvement with salana Solana's Evolution this sort of survey of ZK being used and and other kinds of cryptography being used in the Solano ecosystem and some of your ideas about what could be coming next yeah thanks for that of course thanks for having me thank you Chun as always yeah excited to be here someone someone's got to be the the unruly American in a podcast so I'm glad we had you uh I want to say

thank you to the podcast team Rachel Henrik and Tanya and to our listeners thanks for listening [Music]