Episode 397 - Dev Ojha on the Osmosis Story and his Return to Privacy

Transcript

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 online. This week, Garham and I catch up with Dave Oja, co-founder of Osmosis and a longtime ZK researcher. We discuss the story of Osmosis since we last had him on in 2021, as well as the role of Osmosis in the Cosmos ecosystem. An ecosystem which, like many that we've covered over the years, has gone through

a boom and a bust. It's introduced pioneering technologies but also spedrun experiments in proof ofstake crosschain communication and inspired truly decentralized communities to emerge. Today the network is significantly smaller than it was when we last discussed it. But it's left an everlasting mark on the shape of today's technology and architectures. So join us for this recap. We then move on to Dave's new work on Zcash and privacy technologies in general. He and the group he was working with made significant contributions to ZK research in the 20192020 era but stepped aside after this to work primarily on osmosis. Now he's back in the fold tackling the new challenges and

the problems to solve in the recently reinvigorated decentralized privacy space. We are happy to have him back. Before we kick off, three quick points. First, ZK Summit is coming back with its 14th edition. Applications to attend are now open. It will be on May 7th in Rome. So, I hope to see you there. Second, ZK Mesh Plus. Our paid subscription tier is rolling. We have released a few bonus clips over there as well as some of our exclusive research reports. If you want to get access, please join the paid tier. Third, if you're looking for other ways to support the show, we have donation accounts visible in the footer of our website. Support definitely goes a long way. So, if you do decide to

donate, thank you in advance. That's all for now. Links are in the show notes. Now, here is our interview with Dave Oja. Today, GMO and I are here with Dave Oja. Welcome back to the show, Dave. >> Hey, thanks for having me back on. It's been crazy. I think last time was day after Osmosis launch >> 2021. We looked at this. It's so it was July 2021. We're now talking almost what is it 5 years later. like almost um so this will be very much a catchup episode. I think we want to hear more about what's you know what's the story of Osmosis since we had you on speaking about that. I do want to highlight one other episode that you did before that

you were in Aleandro's group. You were doing like hardcore ZK systems design research papers and that's how I first met you. You had been part of this paper fractal. Were you part of other work back then maybe people might be familiar with? I was part of Aurora Stark work. Yeah. >> And then I helped out Marlin and a lot of the libraries like Libby and Arcworks. >> Very cool. So yeah, I think people are familiar with your work if they've been listening since then. Then you switched over to Osmosis. Yeah, I think what we should do with this episode is do a catchup and I know that you've been getting more involved once again with ZK and privacy. So we'll definitely have to touch on that. Uh I want to say hi GMO.

How's it going? >> What's up? doing well over here. >> How do you guys know each other? >> How do we know each other? >> Uh Tun introed us right after Osmosis launched at >> Oh, that's right. >> I I just unpinned our Telegram chat trading CFMM for ZK. I had it pinned on my top Telegram for 5 years. >> Wait, that's crazy. I totally forgot about that. >> Unpinned it this morning. >> You were just like, "Oh yeah, this guy sucks. Never mind. I'm out." Honestly, fair enough. I'm with you. Like, totally understand. I get it.

>> Old research. Old research. >> That's right. I remember I think we met at the Osmosis, was it a launch party that you had? >> Yeah, >> I think that's what it was. And then we just spent like the two hours or and a half hours talking to each other or whatever. >> Yeah. Great times. >> Yeah. Yeah. Back in the days. That's crazy. >> Nice. All right. So Dave, I think we are due for catchup. When you were on last, you were launching Osmosis. Osmosis was and maybe still is one of the key pieces of the cosmos ecosystem. However, the cosmos ecosystem, especially of late, has, you know, changed, has shrank. It's it's a different world. But I want to go back in time to 2021 and sort of talk

about that arc that Osmos has had because Yeah, it's been I'm sure it's been a wild ride. >> Yeah, it's been awesome. What was kind of cool is I think this is a rare moment where kind of everything you've done connects of like both the fractal work osmosis stuff at like being here day after launch and where we are today. >> So I'm kind of excited to go into it. >> Cool. >> So we launched Osmosis. Part of the thinking or belief was blockchain had three problems at the time. Scalability, privacy, and like proof of stake. We had at the time roughly believed we we kind of mentally solved proof of stake and we know exactly what to do for scalability.

It's just a matter of doing those. on scalability. I actually agree on proof of stake. I think there's one big idea that happened since 2021 which was using like DAG based proof of stake or people now call it MCP stuff. >> Uh but other than that I also think that proof of stake was the shape of was roughly known in 2021 >> but privacy was not at all. >> Yeah. >> So and privacy we we at the time broke into two sub components mev and anonymity. So it was privacy at the cos like this consensus level like where validators basically knowing ahead or people being able to frontr run this was a form of privacy you would want and then there was the privacy for the user which is sort of the way we often

understand privacy in the blockchain context. Yeah. >> Yeah. So maybe today I'd use slightly different words but that's how we broke down the time. Okay. So like okay we want to solve let's first solve ME stuff but solving ME is not a product a DEX is a product this leads us to app chain osmosis >> and then we always tell ourselves once we get the like me decks solid and you can surround it we'll switch back to privacy stuff >> so when we talked first we just launched >> the decks >> just launched the decks >> and then we were actively working on this threshold decryption scheme.

>> Yes. So the idea is that you block me by or you block many class me by not letting validators see a user's transaction prior to being available for execution. Like you couldn't reorder or censor someone based off content because you can't see it. >> So if you you could not up front someone on a trade. >> But was that part of that initial build or was that something like an idea at the time? >> We didn't launch with it. We wrote papers on it and built libraries for it with in collaboration with the NMA. >> Okay. And then we also designed a change to Tenderment to kind of you could think of it as the first MCP like thing. We got this implemented through tenderint but there was a bunch of like kind of a classic Cosmos problem is getting things

merged and like ecosystem adopted takes way longer than just writing code takes. >> So we had this all done I'd say within four months of launch codewise but then this bunch of security stuff and getting this to an ecosystem standard took like two years of delay. M >> so uh at some point we kind of depprioritize this threshold decryption part and focus on other me solutions and instead like go for this path of like well but while we launched Osmosis we actually hit a heavy PMF of this DeFi sum it's like DeFi season back then 2021 and then all these Cosmos coins that people were really excited about like Akos DVPN Iris they they had never had a market before and now getting a DEX

makes their product actually usable So there's like kind of glorious surge of oh people started paying for like cloud computers with aos people could like acquire Sentinel to do a decentralized VPN and trade atom and whatnot. So we we get a lot of demand on this like DeFi side and like start really focusing how do we build a DeFi ecosystem? How do you optimize stable swaps? This is 2021. Yeah. This is also like bull run moment like 2021 2022 and Cosmos all of a sudden like really with the launch of Osmosis becomes what it said it was going to be which was this interchain network of different chains right cuz before that they were all standalone

maybe they existed on like centralized exchanges but they weren't actually attached to each other as far as I know until really you guys launched. Yeah, we're the face of the first usage of IBC. Yeah. And at the time I And I actually still stand by this. IBC is your best bridge of you go from multi-our delays on other websites to now you can deposit, withdraw, integrate in one site, the whole thing is done in 10 seconds. >> Yeah, >> it's way less scary, etc. So now you get this like ecosystem of chains that work together. And then Terra ends up being this killer use case. This is before, but this is like looking back you can see it like a slow motion crash coming.

But at the time, >> people believed that Terra >> was real and could do what it said it could do, not knowing that there was this like fatal flaw that, you know, could even accidentally be triggered, which caused a full cascade of a stable coin. So stable coins, something supposed to be pegged to the dollar, >> fully decreasing in value and losing kind of everything. and then tanking in a way everything attached to it. So yeah, tell tell me about that experience. Actually, we've never we've never done a like a recap on that.

>> So we didn't really appreciate the fact that we were caught up in the US like carry trade somehow. >> Like I don't it's a bull market. You're like oh the product's great. It's like numbers going up, volume's going up. And we didn't we never stopped back to think and stared at why is there so much US volume? what is the source? And it turned out that this ended up being the best colllocated deck because we're not on there's dexes on on the uh like astroport which had high volumes but then uh and there's dexes on other chains like they have this ETH bridge where there's this curve 4 pool and I don't know if people remember that like the whole DPEG started when Terara

pulled liquidity to switch pools. >> Mhm. So, there's stuff on Ethereum and then there was stuff on Avalanche. >> This one's a little bit memory holds. I don't quite remember the details, but there's some like den box trade or juice box. I forget the name of it where there's high a high leverage trading strategy that went through avalanche somehow for for yield farming. >> I this don't remember actually. Oh man. >> Okay. It's part of the part of the carry trade. Um >> I believe that. >> Well, so we thought so we look at BS. thought those were the sources but it turned out kind of empirically oh no there's actually a lot that is uh was going through osmosis and that was part of the huge like pump in price and then like volume numbers at the time so USD

collapses this almost like existential so there's a couple existential things that happened at the same time okay one there's no other stable coin in Cosmos it was just USD >> why well we actually were trying to get an Ethereum bridge at the time. What happened is I carried a tower of bridge security. I still do. >> Yeah. I remember you guys did a runoff, right? You like were interviewing different bridge companies and like >> who did you end up landing on? >> We used Axar at the time and the reason uh it went through governance, but I was actually really proud of this as a reason. What happened is uh there's four competitors. There's Wormhole, but Wormhole just had a $200 million hack on Salana.

>> Yeah. >> There was this team called Gravity Bridge, Axelar, and Nomad. James Press which is project >> which also later got a big hack that tanked some other networks. Yeah. >> Yeah. So we talked about all security and like guys the bridge is a single point of failure. It you cannot have any system where a single bug or single hack can make you go from like go to zero especially if a lot of money in it. >> This is pre fail I guess right. Oh wow. So this is like your focus had been on the bridging. And you were like, "The bridging's got to be great."

>> And meanwhile, you had this like disaster token in the heart of what you were doing. Oi. >> Like, so like maybe like four weeks pre- crash, we talked about this. We're like, "Okay, we need to get the get the ETH bridge thing finalized, get everyone to like finalize this, >> but it hadn't quite happened." >> No, we're cuz we also didn't think the Terara thing with US was as pressing as it was. We thought, "Oh, it's going to maybe there's an issue and it could be a slight dagg this year." not like this is imminent days matter. So we talked to each of things of like hey you need to ensure you can't go to zero what's your strategy for rate limits or something else is like yeah we agree we have rate limits in audit right now >> awesome at the time all the other

projects had some varying answer of no rate limits allowed >> um so Axar ended up winning this vote which is great because there's no way axar can have a bug and a user goes to zero >> but this happened like two weeks too slowly And so there's no other stable in Cosmos. >> And so people had to switch. I mean I there was probably like a mad dash like the the volume at the moment of the terror crash. What was it like for you guys in Osmosis at that moment? >> So what happened is all of the Luna assets were internal assets. So Osmosis was the closest decks that had any other assets you can trade against any of the liquidity. So we actually there's two serious bugs Luna as the crash is

happening. One was that there's a Luna runaway mint cuz the way it worked was that you could like burn US to get Luna at the peg price. >> Yeah. >> Well, the amount of Luna supply was massively skyrocketing cuz it grew pretty slowly and it was constantly more like Luna US exchange, but it crashed quickly and tons of USD tried to get to Luna to sell on a sex. >> Wow. >> So, these proof ofstake security of Luna was in danger. So, we were actually worried about could this be the first twothirds takeover of a blockchain network ever.

>> Oh my god. >> Did it happen? >> There was a bunch of Telegram chats where the the Luna validators are basically like point we point this out as a problem and Luna Valor is like, "Okay, we will halt this chain and upgrade to fix this." So they they choose to upgrade. So they organize a hard fork and halt and just stop allowing staking. >> Yeah. >> And so Luna Classic had that as a thing for years. Then there was a second thing that happened. I'm trying to remember what the cause was, but they basically halted a second time. I think it was to disable the mint function entirely cuz they were like, "Guys, this this is like not helping anyone. Let's just we lost this. Let's not let more US at like 3 cents become Luna." And so at that

upgrade, they close it. And before they relaunch, somebody this is not any from Osmosis points out that like guys what about the Osmosis decks? You cannot screw over all the DEX users but like more exit liquidity. So they freeze the IVC channels as well. >> So that was a huge uh a huge relief for us. It was shocking. We didn't help coordinate that at all. They did internally and so from that moment we had a little bit more like calm. >> Okay. >> But on our end what do we do? Well okay this is happening. What can we do? we can freeze these pools, these like US and Luna pools. So, if you're an LP, at least you don't go fully to zero.

>> So, we build this logic and get and then like in an emergency. Yeah. And then get on calls with the community like we have tons of voice calls and recordings to show everyone like, "Hey guys, we need to do an emergency hard fork in order to have protect these LPS while Terra's halted." Then we do like basically several series of audits in public on calls with people. So we get a bunch of the Cosmos devs and like SDK devs together to help look at this and get an upgrade deployed in like sub 24 hours. But this is abnormal. This has never happened in like kind of the history of Cosmos. There's never a security bug in that a token droid security bug either.

So nothing merited like emergency upgrade, >> but this did. >> Yeah. >> Interesting. This, by the way, for listeners who aren't familiar with the history here, Osmosis launches in 2021. This happens in like May 2022, I think, or June. Yeah, sounds right to me. >> Something like that. >> And following that, there was the Nomad hack. There's like side note, there was like the tornado cash case comes up and then there was the FTX at the end of 2022. Yeah. So, those are like the the big hits, but this Terra crash was the first. It was the first and like it killed >> major VCs and trading groups.

>> People went to jail and were caught in Montenegro and put Oh no, >> they were extradited to the I mean there's so much this is like big big disaster for a nation. >> I feel like this is like a good thing to go ask your like local chat GPT provider you know about cuz I feel like it's like a weird moment in history that is like kind of hard to synthesize. Nobody has done it super well, but surely >> weirdly, I mean, that cascade of 2022, 2023 was sort of chill, 2024 was sort of chill, and really it was chill until like end of 2025 when there's another cascade that we have never explored on this show, this October 10 >> 1010. Yeah.

>> Yeah. This like and and it caused a whole cascade of activity as well. But in both cases, I guess it was like overinflated ecosystems in a way, overinflated dreams getting totally crushed. I don't know. Sometimes people think it's malicious. I mean, I know with the Terra conversation, there was like claims that it was done on purpose. Same with the 1010 thing, but I mean, we don't know, but okay. So, Osmosis at this time, so you go through this like like emergency upgrade, and you're still building osmosis. You're still looking to like Yeah. Our strategy pivots then where before this we were really focusing on like growing Cosmos crosschain standards and a lot of security stuff super food staking rate

limits etc. we still take security seriously of like we okay rate limits even more important maybe we should add those in more areas like with like for LPS or something which I can make a tangent I think that's still separately a thing that should be done for all lending markets and big pools like hey everyone should always be able to like exit their D5 position immediately but hey if I have a lending market with 100 million USDC maybe don't allow 100 million USDC to be withdrawn immediately if 100 million wants to pull out deexpose lending it, but maybe have a cue of like we have a 24-hour period to like check was there a security bug, did North Korea break into someone or is like real user activity cuz that's basically what bridge rate limits are there that for bridges you should have

that throughout DeFi as a safety mechanism. >> We start thinking about a lot of that. But at the after terror crash, we pivot to okay, there's unique opportunity here to terra how to build a massive cosmos ecosystem. At the time, I would argue that this was bigger than the salon ecosystem 2021. >> Um, yeah, >> this was before the whole like Saber, the Macalo twins are 50 of the projects thing became public. >> Also, wasn't there like after Terara and Exodus to Salana? >> The Terara devs went ter they did a great job of building developer ecosystem. Shockingly, they went to three camp I'd say there's three to four camps them. One camp was just glorified anchor rappers that were basically all

kind of grift there. Then there's a lot of people who went to want to cosmos and then they switched to either coming on to Osmosis. We did a lot for this >> or starting their own Cosmos app chain or going to like Juno or something. And then the third campa went to Salana. >> I don't think we saw anyone from Terara go to Ethereum, but someone could correct me on that. Uh >> I'm sure there's people who did, but just like in in terms of large flows. Yeah.

>> Yeah, that wasn't the pattern. >> Makes sense. >> So then in a nutshell, we go like we try to make it in very integrated DeFi cosmos strategy for a while. Then we run into a problem that there's not there's not many new app chains coming up that don't try to be their own decks. So Celestria launched and one very cool technology stuff, >> but there wasn't too many after Celestia. After Celestia, we got like massive run up in activity for osmosis and the like. >> So we have a problem to confront. Where is demand going to be? >> Before Celestia, we're evaluating very carefully. Do we want to do PERPS or not? And just fundamentally we decide PERPS are a winner take all market.

There exists Hyperlquid, Drift, etc. Hyperlink was probably really on to this with strategy, but we disagree fundamentally on like you need to be max decentralization. We don't want to play the per game. I just there's many things you can imagine your life doing. I'm inspired by privacy. I'm not inspired by making building per game I want to play with my life. >> Fair enough. >> So like we're conversants. What do we do problem? So we start building this very cosmoscentric or appchaincentric model of the world. We realized users don't want to think about chains, but there's this like weird structure in crypto where you are forced to >> at that time kind of the only big crosschain wallet was Kepler. Phantom had not added any EVM support. MetaMask

didn't add Salana support. That was very recent. >> Um, so none of your big wallets really cross trade. I mean, technically MetaMask supported the roll-ups, but it's like probably one of the worst integrations ever. So like, okay, users want to just trade assets in spot. They don't care what chain they're on. So how do we make a UI that can let you come in with any wallet and then trade across every single every single chain and hold your funds on all these chains? So this is the start of this idea called Polaris. Uh we're going to integrate all these bridges, all these dexes. And once once we get Polaris to succeed, we'll do a lot of work to bring the flow onto Osmosis.

>> Is Polaris a wallet? >> So this is a part of the problem with Polaris. It was very hard to succinctly describe >> cuz it is in some large sense a wallet that you'd have. It would be this desktop website where it's kind of your exch your exchange and custody solution. >> But would you still use like Kepler as your signing wallet or something? >> Yes, exactly. This ended up being an idea a lot of people converged on with the biggest difference being how difference of how you do custody and how do you clear the swaps where we were pretty clear on all these intent systems don't have any liquidity all these projects are using their own cap table or like trying to purchase like rent liquidity from wintermute or market makers to be on there rather than yeah they're being act normal people feeling

as intense and we actually saw this in practice until Zcash prompted last year there was very little fillers for all these intent systems across chains. But then after Zcash like near intense started getting a lot more like natur more organic filling like more participants come in. It was just hard to get the volume you need to onboard market makers to a new system especially when it's unfamiliar to them. >> So we start building that out and come around May last year I I start taking like a lot more charge in this and try to make this much more privacydirected. So I set us up on, hey, if we're building this wallet on your behalf, you come with Kepler, come with Ethereum, why don't we keep your funds privately on Namata.

>> Okay. >> And then we if so, we can hold anything on Namata and then you can do a swap in our UI just like a normal swap and it'll bridge across to Salana, do a swap there and bridge back. So to do that we'd you we built this like very complex and advanced NPC wallet system where the NPC wallet would fully abstract the shielded sync. We'd have the MPC committee like in sort of a multi-party way. So actually I'll explain maybe I'll explain sync it own bucket. So solve the hard part of privacy and then we solve usability because we'd have this committee do all the bridging and swaps on your behalf but you as a user did one signature from your existing wallet with some like human readable text and the MPC wallet would affect in the back end

would follow the text you laid out. So it's like obeys the non-custodial kind of rules. So this would let you kind of have one an anity system that could do lending do private swaps across all these chains which is great. That's 8020 of kind of all crypto usage. >> So was the privacy part the namada system was that where the private >> things would sit and then the rest would kind of bridge out. Okay. >> So we inherited a few problems though. We inherited a lot of dependency problems of we had to depend on like Namada for upgrades and a lot of third party bridges and these two problems ended up being kind of fatal for us where like a week before launch one of the bridges that was critical to this

working cuz the market was Salana this wasn't making commercial sense for them and it would only make commercial sense if we like massively succeeded we couldn't guarantee this so like couldn't get a proper agreement so that didn't work out and we there was like a upgrade Namatada that was like just couldn't get couldn't get shipped >> and actually at this point is Namata even maintained like I think it's not as the focus of the company >> yeah today Namata is not um at the time this wasn't like announced and stuff so what you see where things go it kind of makes sense cuz like >> sure >> a big problem you have is like when's in decline you just can't bet on all these dependencies being there this whole app chain thing is really exciting if you have and works great if you have a lot

of dependencyless infra infrastructure. >> Mhm. >> So that's one of the great things about IBC. It's permissionless. You don't have this problem. >> But all most of the Cosmos bridges from like Salana to Cosmos, we're not obvious like they're multi- like. So you have to have uh that's a big dependency you that has to work. >> Yeah. What has happened to the IBC? like h there was a time where like a lot of different projects and chain like Ethereum was going to have an IBC or like they were trying to push this idea that you would be able to have the IBC on the Ethereum side as well. Like would you say that the tide has changed on that now? Like have people just moved away from IBC completely and moved more towards these like what are they called like the these bridges that are me now

very common where you just have locked on both sides and you have like a a market mechanism to like keep people honest. >> Yeah. Most of these bridges have ended up devolving to being like very tiny multi-igs and then you do you have a market maker type system to make it fast to transact better than finality. The market maker system on top is fantastic. The fact that the base has ended up being a multisig basically due to BD and brand reasons is kind of depressing. That's the exact opposite should be the case. >> Yeah. So I mean it sounds like theoretically or like your heart is still with IBC but I want to know about the reality of IBC. I actually I think the heart is still in IBC and I think this is like a little bit of skill issue like this is just no one owned the go to

market well of it really life is like the world's full of path dependencies and I don't think the path dependency of rollup to roll of communication or roll up to Ethereum is like >> good I think it's a path dependency thing not a fundamental thing with IBC >> okay >> so where you are today I don't know how you directly win other than like permissionless infrastructure is great and should win if you have a lot of permissionless chains >> now. It's kind of we're not seeing a lot of permissionless chains. You're seeing isolated like corpo chains. Uh a couple big rollups Salana. Yeah. Tempo corpo chain. So then these corpo chains are going to be perfectly fied I think with like centralized agreements to a multisig. They might even prefer that.

>> But unfortunately the multisig bridges won the branding. So now it's somehow a prestigious brand. >> So as a corporal chain you're going to prefer that. What? >> Yeah. I I would argue like layer zero is considered like the good bread for a lot of like USDTs and stuff. >> Mhm. >> Like people want it for the bread >> or a cross. >> Yeah. Oh, so a cross is a fast bridging thing. It's a market mechanism on top that doesn't care what's underneath. That's actually fantastic. That can apply to anything. Well, it just like is independent of the underlying layer, right? There's like like the market thing is >> I'll pay some extra fee to do the quick bridging but somebody else eats both the time risk and also potentially like the bridging risk depending on how you see it of the underlying layer actually

bridging the things separately. You also can get smart about how you do certain types of bridging and all sorts of nonsense and that saves you some money and then you batch a bunch of stuff together and all sorts of stuff but for the user the like highle market mechanism is really nice. It's like look man I don't care if I'm paying 10 bits or something on my like bridge I just want it now. Yeah, the UX of all needs to be fast. Like the UX of too many things is crypto so slow. So I actually think across is great. I wish they did we could do more things to make like the fees lower, but this is like Yeah, it's absolutely the right thing. Get this uh get Marcus make us fast. I want to sort of wrap up the story of osmosis because I know that you've recently been I mean

there's a there's news about it but also there's new work that you've been doing but I also want to just mention something else kind of that's happening I think in in tandem which is like the terra breakdown the 2023 shift you're sort of focused on a new system cosmos does get smaller and smaller there's a Celestia boost like Celestia launches and it kind of like rekindles a lot of excitement about it but I remember the narrative after that being like, well, is Cosmos Hub really the center anymore? Is Celestia the center? Would a stable coin thing be the center? Like, where is the center of Cosmos? Where, you know, where where should it live? So, kind of like what is your postmortem on Cosmos?

Maybe it's not over, but like where are you at looking at where the ecosystem has landed so far? Um, so maybe around when Skip took over, I was pushing really hard that like look, privacy is the thing that matters most. >> There's this opportunity to make Cosmos this like the center of privacy in all privacy apps >> cuz it would fit like the spirit of it in a lot of ways >> and in many ways like a lot of the privacy train privacy work in crypto aside aside from Zcash was happening in Cosmos like whether like you or not secret network Penumbra was in Cosmos.

>> Penumbra, fantastic. Henry, genius guy. He's come to comes to Cosmos, makes his own chain to fix like all of the problems that Shield would think he uh he'd find and get the wallet UX right. A lot's happening like here. And this is like pre- rail gun privacy pools. Okay, I guess we're having this conversation. Rail gun has launched. >> I don't remember the dates of privacy pools, but like I I think we could really do a great job. Got the talent. like this is the thing I'm most passionate about. Details happened. Couldn't convince along those lines. >> I still think this is super valuable.

There's several people doing it. Like Aztec, Starourt just announced a lot of privacy work. >> On the T side, there's a lot of people, but like ZK side, I think these are the two main ones. >> Well, and like Zcash, of course. >> Sure. >> So, where is Osmosis at today? >> So, today we're basically Cosmos dying. Like I think aside from Osmosis, almost every chain has announced they're shutting down or like switching to another ecosystem. >> End of an era. >> Neutron announced they're shutting down this morning, which was the kind of the home of a lot of the permissionless Cosmos devs.

>> Yeah. >> Wow. >> So Osmosis is I think in order to save where Cosmos is, what we need to do is merge Osmosis in the hub. Get the get the decks here. Get the permissionless Cosmos ecosystem happening here. skip is rather cosmos labs kind of ICF is trying to get like enterprises to use the cosmos stack and like I don't try to secure very large BD deals and like sport agreements but in order for that to funnel to like Cosmos tokens and the ecosystem you have to have a way to kind of capture value from that beyond just an enterprise deal so we've kind of seen the the reliable way to get high amounts of liquidity and volume flowing through to generate revenue is use the decks. Uh

so I think they should clear thing to do I think is acquire the decks and then we rebuild the cosmos ecosystem on the cosmos hub now it's permissionless and the decks will be colloccated it again and like research this from that >> and hopefully once that starts succeeding initially like bring in start doing privacy stuff back on the L1 >> okay >> there's a gap in still of getting a good contracting environment with like native privacy a lot of people are filling it in now but I still think it's a gap from where we need to be and where we are today >> interesting >> very quick question before we like go too far is where are the actual chains that are currently leaving the cosplay ecosystem and devs going out of curiosity.

>> Devs is kind of an open question. >> I think it's like kind of in sporadic or like it's no secret that a lot of devs have just left crypto. So I wouldn't say there's like a clear nexus of how they've moved the dev side chains. I'd say a lot a lot shut down or a lot went to being on one of the Ethereum rollups or Salana directly >> more more so would be Salana. >> Makes sense. Yeah. It feels like Salana is a more aligned of the two ecosystems here and has been for a long time. >> Yeah. And I think it's kind of natural when you have a bare market you need a lot of consolidation like okay let's not be an isolated app chain where it's hard for capital to flow into where one of the bets of cosmos is that you could

have had a lot of capital flowing into new chains easily and >> when I think of like logic brain this seems like it still feels actually correct but just we look at the market reality that's not been the case a lot of a lot of like enterprises want specific deals per chain >> getting a custodian for every single chain is a lot of work It's not a free operation even though it should be explorers are getting a lot better especially post AI >> possibly that too. >> Okay. Anyways, now we can find all these. Sorry, I didn't mean to hijack. >> Yeah. So, let's let's now switch over to the work that you've been doing on privacy. This we kind of at the start of this episode we were talking about how the episodes that we did in the past, Fractal and then Osmosis launch. Yeah.

>> Like and actually we did a bunch of we did this privacy in Cosmos series of videos. I mean, I remember you guys talking about a lot of the ideas for ZK in IBC or ZK and no, it wasn't in IBC, ZK in the kind of validator stack and all these things. So, kind of how are you tying all of this together and is Polaris still a thing for you or are you usually shifting focus? >> No. So, okay, I can think I wrap a through line of Polaris to what I'm doing today. >> Okay. >> Hey, so with Polaris, the dependence risk was too much. Shut down and I was like, okay, how do we bet on privacy? I had been talking to a lot of the Solana projects in parallel, all the a lot of the wallets, the devs, and this is post

Zcash pump. >> I talked to all these people precash pump and everyone's like, "Oh, privacy is cute. Doesn't really matter." >> Yeah, >> I talked to them during a Zcash pump. Perfect. We want privacy now. It's part of our 2026 road map. Um, so then I we designed like the shielded system for Salana that would solve there's this big problem in privacy UX which I should talk about in something called shield sync. We designed a system that would eliminate this on Salana and actually got it like built out and demoed. But there's a really big problem I want. I think the dream that like really changes the world for where we need to be, especially with like AI, global surveillance, and whatnot, is to have privacy with like no global master key that could like somehow be activated to

decrypt past transactions. >> Mhm. >> And in order to make that happen, I think you have to go out about it in a very maximally decentralized way. >> And so I realized like I didn't think launching a new thing on Salana is the correct strategy for success. So because that ended up dropping that and therefore this Polaris line of work but the code is actually ready for that. So that then takes me to what is the most decentralized thing that has privacy at its root and a massive community that has that would not compromise that gets you to like Zcash. >> So then I've been what I've been doing lately last few months is putting huge amount of energy into along with other wonderful folks in Zcash. think talked to Shawn Vo recently. Yeah. How do we get Zcash to like scale and be

incredibly usable? Well, and so I focusing on both of those problems or I guess focus on three prompts scale, usability, and like postquous of it. And the usability might have a different definition for me than what may be expected. >> What are you doing with Zcash though? Have you joined like the teams? Are you contributing? What's happening? So what I've doing is I've found started a new company that's effectively uh I mean if socially sanctioned a core contributor to Zcash. My my general goal is make like privacy win do whatever it takes to help like kind of privacy anywhere win but I'm focusing almost all the time on Zcash directly.

>> Okay. Why did you choose Zcash and not one of the other privacy systems like Aztec just launched or >> Yeah, it's a good question. So I I love Aztec folks like like Zach a friend Zach and Ariel brilliant at doing stuff for incredible amount of time. I always think about Zach as like huge for inspiration for just like grit of >> for sure. >> But doing Aztec unwaveringly I think like seven eight years >> and pro privacy the whole time when everyone got scared and cold feet >> he stayed with it.

>> Yep. >> Yeah. Awesome guy. I think that the thing that matters to really win is to get a lot of the payments the payment flows to work and build everything on top of that. >> And then I also think that you either need to be colloccated where all the users are or so like making a new system is really hard >> or you have to be something really pure. So, one of the things that like is always shocking to me is I was really inspired by some parts of Zcash of like when Zcash pumped, privacy everywhere became the thing >> and I think that's only possible cuz we what winning looks like is not like hey all of payments have this like privacy property that's key for like freedom winning to me looks like we've inspired

like self-owning your data across the entire internet like you have local LLMs all your data is only owned by you never shared there's we start getting people to like just think about metadata leakage in everything they do self-signing. >> I don't see a path for anything that's not super pure winning to inspire that wide movement. >> Whereas, so I think like if Zcash wins or Zcash like really improves a ton, this has this ability to inspire a lot outside of just crypto. >> Whereas that it feels like very programmable and too too finance heavy.

it it's a muddled message that doesn't spread as far. Like I'm I'm always so shocked at like how successful and impactful Bitcoin is despite like look you can barely even pay on this. The thing they got right was just that like the security of it it's not had had an incident >> and its message is very pure. >> Yeah. Even if behind the scenes there's like a lot of random miners and blah blah blah. a lot of stuff, >> right? But it is like a messiahic message, right? It's just like brought forth from the Messiah, you know, whatever.

>> And I also think that like it's really key that we get crypto native currencies to have this because if you want a privacy property that's really big, it should be something that's like kind of digitally enforced similar to like in order to be inspiring, it must be like unwavering. You can't have a bunch of asterisks of like you can your privacy can be broken if XYZ conditions are met. >> That's impossible for anything bridged from the real like like real world it's on digital native like if there's if it's you have digital gold or if you have like Paxos gold it doesn't matter anyone tells you whatever system magical systems you build Paxos can like say you must dox everything or else uh you cannot use this. H so you have to get a digital first system to like be

inspiring and make this whole movement. >> Okay. Yeah. So, so I guess maybe here's a a bunch of questions is number one is where is Zcash today? How usable is it? Are people using I remember there was a bit of a runup on it. There's a whole thing. >> Yeah. >> There's also like a wallet split that's just happened. >> Oh, interesting. I didn't hear about this. I mean >> Oh, man. And then number two is okay, what what is your dream for for Zeke? you know, like let's say you got today to make the the the philosophical things come true. You wave a wand and things happen. What would that be? But let's start first with where the hell is Zcash today? I actually have no idea. I mean, no clue like whatsoever right now. Okay, so where are we today with Zcash? Number

one problem is kind of usability of this. So you probably talked about a couple times in the podcast like why price suck to use due this thing called shielded sync. I mean, specifically the Zcash model had this like transparent and shielded versions. You had to sort of move and decide where your tokens would live. >> And I think what I understood from >> Sean and we we just recently did kind of this history of those wallets and systems. So, I don't think I I would actually point people to that episode if they want to hear that. But the idea that like every time you wanted to sync your wallet, you'd have to kind of download the entire >> Yeah. chain and you'd have to go through all you're running all of the

transactions that have happened in the past. Meaning like and and I've experienced it where you're trying to open up your wallet you haven't used in a while and it takes forever >> for it to be usable. Like you can see it like sinking and you're like whoa this is like hours and hours and hours of time. So that's like been a big issue with that particular system. Now, other privacy systems have other trade-offs, and some of them have been faster, and some of them maybe like when they just launched, they just didn't have that much activity, so they didn't have that kind of delay. They didn't have these long sync times. I mean, I I don't think Aztec's built in the same way. So, I don't know that its wallets would have the same issue.

>> If you wanted the same privacy guarantee, it would. Okay. >> Uh, Aztec does this trick that's actually it's pretty clever for if I know I'm talking to you or I've paid you before and like somehow my wallet can figure this out. Okay. >> Like I don't know. I imagine I have an address book. I paste this in. There's a quick way to resync. >> Now, I'm unfamiliar with how does this address book work in practice? Like like is there a public directory of all the addresses and it's just not that many so it's quick? Or you enter it? I actually don't know the answer to that. But once I know your address, I can then find all payments between me and you rather easily, >> which is a very cool property.

>> The way they do it breaks privacy for now, but can be fixed with this technique called private information retrieval, which is actually kind of pretty centric to what I want to do and have been doing >> P. Yeah, >> I think we're actually planning an episode on that particular topic. So, >> perfect. Yeah, it's one of the most exciting things where it's shockingly usable. We've been doing it for things at more than a thousandx the size of all Zcash transactions to date and it works great. >> Cool. >> Very cool. >> So the problem in a nutshell is basically you have to download all these things all these transactions. So we need to change the payment flows and in order for you to no longer need to download all every transaction on the chain. But we also need you to be a we need a second property where I could fix

the first one cuz what are we solving for? If I pay you, how do you figure you got paid? >> And there's like two kind of somewhat simple answers. Because we're a private payment system, so no one knows who it's from, who it's to, or how much it is. Easy answer one, you just you download every single transaction in the chain and check if it's a payment to you. >> But that doesn't scale for obvious reasons. >> Second easy answer is, well, why don't I just tell you I paid you? Like I can just send you a signal message that hey Anna I sent you $10 for coffee and that would solve this.

>> Yep. >> But there's a problem. >> Well that's not guaranteed that you did it right. >> No it can be guaranteed. Like I can send a link that you click your wallet opens the link and it it's for sure received. >> Okay. Okay. >> The problem you get in there is two big problems. One is what if this like communication channel fails? My phone breaks. Your phone breaks. Do we have money lost in the middle? >> You have to be there to receive it somehow. >> Yeah. Or you have a central server who receives it, which is how every messaging app works, and you try to do a bunch of work to hide the data you pulled >> uh from the central server.

>> Okay. Second big problem when you want to, let's say you've lost everything, like your phone's broken and you have your 12word pneummonic. You want to go, you know, put it into a new phone or a new laptop and go recover your funds. Well, if you find out you got paid off of signal messages or some messaging app in the background, how do you get your money back? It's no longer 12 words everything. >> Okay, >> so we fix both of these problems using this private information retrieval trick. The idea is actually quite similar to Aztec which is why I familiar with it where we segment the problem into two camps. How do we do the first payment the first time I ever pay you and when you and build a way for that to be recoverable. So when you recover from

your pneummonic we've built a system so that you can figure out you got a first payment from me. Second thing we do is for all subsequent payments, if we've talked to each other once in the past, we can kind of agree on a key and or we can agree on a kind of infinite sequence of random numbers. I call these tags. And for my second payment to you, I'll use like the second tag in the sequence. Third payment, I'll use third tag. Your wallet when recovering can just know to look for this tag because it knows what these what the tag numbers will be but it needs to find is there any transaction on the blockchain with this tag. So only me and you don't know what numbers correspond to us but then any

node can index one of the tags it's seen >> and we build a way to privately ask that question. That way you private ask question is P. Okay, this isn't what I'm just trying to remember because like we talked about this like pruning the nullifier set that Ian and and Sean have worked on. That's something else, right? >> That's an unrelated. So that's a cool problem that's like longterm. I think that's further out in the problem stack reusability, but it's fantastic cuz the there's another problem in Zcash or like systems. Every single transaction you do makes the entire state of the blockchain bigger >> with these nullifiers which are very small but yeah eventually these things can add up.

>> Exactly. >> A billion nullifiers you get you get to a billion or say you can get to two billion nullifiers with with a billion transactions. So that would cost you 64 gigabytes of data. >> So if we get to hundreds of billions of transactions then we get to very big numbers of database sizes. >> Okay. So we need to long-term solve that as well so our nodes like stay very usable and tachons actually very brilliant solution to this and then we have a second problem which we need to kill the shielded sync we need to kill shielded sync in a way that still lets you recover all your funds.

>> Okay. >> Mhm. So the ladder is where this like P stuff comes in. So you can recover from your 12 words, find all your funds and uh yeah, and still spend money, >> right? >> You don't have to have any of this signal messaging or like you know offchain messaging. >> But are these kind of related cuz like one is about the bloat of the entire system and indexing or like being able to retrieve information and then this one is like these do seem sort of similar. Would you have to adjust the PI in order to like work with the new system?

>> What's funny is this this came up the kind of the first time Sean and I talked about this where Yeah. actually brilliant. So I just got a couple people and no one's like actually asking this connection. So Sean told about like how there's evolving nullifiers. We have to like kind of update the nullifier every epoch. >> Mhm. >> This P thing is very clever, but it does have a big cost. There's a reason this has not been done kind of everywhere. It's that what's happening. It's kind of the server has this database. The client is asking for some entry in the database, but the server and the server has to tell the client this entry >> without knowing what the client asked for.

>> Mhm. >> In order to do that, the server has to do some math operation on every single database element. So, normally if I ask for like I've got a terabyte database, I ask for an entry. It's very quick depending on what like level of abstraction you use. It's like 0 of one or login like some constant amount of time basically here. No, it grows your database size. >> But this the server grows the database size. But the anonymity also gets better the more entries there are. M so for the evolving nullifiers sort of updating nullifier question it's not super clear to me why you wouldn't want to update this every hour other than like it makes a lot of client queries it's a lot it's

like more CPU work uh sorry not even more client queries it's more server work to do a lot of the update proofs >> so it's kind of funny we have these or you can make it really long so you have less client interactions less less bandwidth so we have slightly different usages of kind of time boxing history between this P stuff and these evolving nullifiers. The word is oblivious thinking services. >> Okay, >> so we need to we need to make these two chunks of time the same and make these kind of shared across both systems. >> I'm going to be doing an episode soon on verifiable databases and it's like >> doing kind of information like proving something about a database without like knowing what's in it in a way. And I'm

wondering like in the Zcash specific context, is this a technique that could be used here or is it already happening? >> That one's more like I think if I recall correctly, that one is more like providing a ZK proof of correctness that the thing has been correctly run and like not revealing the underlying data. This is like the opposite case where it's like what you don't trust the person who hosts the database. So in in that one you don't trust the person who to correctly give you the output but you do trust them on the private data. >> Okay.

>> Right. This is the opposite. This is like you trust them to give you the output because you can verify the output in some other way >> but you don't trust them to not like leak what record you are looking at. >> I see. >> Right. Because like just the fact that you are searching for a specific record in this database is leaking information. Makes sense, >> right? Whereas like in the other case it's like more about general queries like SQL queries. >> Nice. >> Okay. So we ended off like how do you we tied off like P connects. Let's make this P database time the same as the oblivious syncing service time to get both goals solved for scaling.

>> And where are we at? Like I know that in Tachion there is I mean I don't know if this nullifier kind of pruning not non-p pruning pruning they said it's not pruning. So I was like, "Okay, okay, okay." Technique to avoid pruning the nullifier set, but actually shrinking it potentially. I I don't know if that's actually going to be included in like in Tachion, but is the PI stuff going to be included? Okay, so this is an evolving target. The dream is end of this year we get we get these evolving nullifier proofs from the Oblivia sync service in and this P stuff um in order to give basically to fully unblock scale. But for scaling, the biggest bottleneck for scaling is you kill off shielded sync that you can now

get your find all your transactions without downloading everything in history because then it's just a straightforward like how do you scale consensus to scale TPS directly. So we're trying to hit like getting several hundred shielded TPS >> on train like this year. So the one of the biggest unlocks is one of the precursors for the oblivious syncing service and all the tachon work is recursive proofs, recursive snarks for everything. >> So that's that's for sure needed and we're we're trying to get like everything together. P O oblivious sync and these evolving nullifiers to get kind of this dream system of no more shielded sync. It's kind of theme is Manhattan project for privacy. In in that interview with Sean though, we also

talked about like what an upgrade would look like for users and he said that like there's parts of Tachon that will really fundamentally change the way these wallets work soon. >> So >> yeah, so it's like I I think this is a pretty major upgrade coming. >> Yeah. So okay, I can tell you a little bit more about this. This comes down to the problem of how do you make the first payment? So I kind of sketched verbally just not maybe not super convincing that if we've paid once before and you can figure out the first payment from me to you >> can then find every subsequent payment.

So now we have to solve how do you find out that I paid you for the first time. >> In this case you want this to be find outable. It's like you're not >> Yeah. I want Dave pays Alice pays Bob. >> Yeah. >> Bob needs to find out that Alice paid him. >> Okay. This is solving for this like signal problem. Yeah. >> Right. like now it's like okay assuming we have the first payment we're good you know solved by this like previous I believe is the name that you had for it >> now the question is okay cool but like sick we still have to find the first payment bro like can we fix that >> okay so here's basically the options that have no risk and then uh okay no risk no risk basically means there's no intermediate state that can like have no

fund that can lead to funds lost then there's things I can like trade off anymore so the first way is Alice wants to pay Bob. Bob Bob sends Alice a dynamic address. So it's an address that's unique to this payment >> and I include like 32 random bytes. So if you're familiar with the I was calling these tags earlier. Bob tells Alice basically pay me with this tag >> and Bob can know to look for that tag of P if he has to recover from pneumonic. So then Alice is a payment at that tag.

Bob's constantly looking for that tag to figure out did he get paid? That's kind of way one where you have a dynamic address or dynamic QR code and we're at a conference, you scan my QR code, I give you a different QR code. It's kind of the idea that flow you can also back to this Alice Bob example where Bob gives Alice dynamic thing. There's also the normal flow of invoicing, right? Like if I invoice a company, I say like here's this like bill, pay me pay me uh $100 and like here's this invoice ID. And it's kind of the same thing of here's this custom data for where I want to get paid.

That's another way to model this. Two mental models, invoices. Second one is sending you a unique address. >> Mhm. >> Second solution is well Bob can recover funds if it's he's talked to Alice before. Okay, we're in a situation where he hasn't talked to Alice. Well, what if there is a third party >> Eve? No, Eve is always the bad one. >> Yeah, I start with R. >> Charlie or something. >> R for relayer. So, this third party Royce Robert who's Raven. >> Raven. Okay. Raven's a cooler name. >> Raven's cooler. >> Okay. Uh, so Bob's talked to Raven before. So, in Bob's address, he also puts Raven's address. Raven's a relayer.

So, Alice wants to pay Bob. Alice hasn't talked to Bob before. But Raven promises to always kind of be online. So Alice talks to Raven or Alice sends like a microp payment Raven called like I don't know a tenth of a cent or something and then says like forward my public key to Bob on the channel Bob can see. >> So Alice talks to Raven. Raven sends a payment to Bob like a microp payment called like again like faction ascent and then Raven sends a proof to Alice that hey I communicated your key to Bob so Bob will be able to recover this >> now go send the full money to Bob >> so now let's say Bob lost his keys and

has to recover from pneummonic well Bob's already solved the Raven communication somehow in the past so Bob will see Alice's public key because Raven told Al um Bob Alice public key. Raven gave a proof to Alice that she did this. So now Bob will then be able to find Alice's payment. >> Mhm. >> Yeah. >> So this is kind of like you have a relayer in your address. Somehow it >> it's playing the it's very similar to me to it's akin to the role of the sync server in like local first software, right? >> Exactly. Oh, that's a brilliant analogy. >> It only keeps track of the deltas and it doesn't matter whether it's online or not. You can switch to any sync server you want and whatever. But the point is like you need some mediation between two nodes which you don't know will be

online constantly. So it's nice to have this as a service. That's that's actually really [ __ ] cool for specifically for like enabling the this like cold start problem. That's like that's brilliant. I was going to say something much dumber which is you could also just have ephemeral data which runs on the blockchain. You know you have like a data availability style layer and then you do this but it's it's it's exactly the same service. this I mean fundamentally the point is here of course is much more general than like it needs to run on the chain itself. >> Oh that's kind of cool. One difference with the chain version that's kind of interesting is you don't need to provide a proof that you have sent the like public key to somebody else is like the chain itself.

>> The fact that there's inclusion in it is enough to provide the proof, right? >> Uh oh, that's true cuz you could have the Raven to Bob communication something not be encrypted and you just see that they sent Alice encrypts the data of her pup key and so she sees that her encrypted data is used. If Raven doesn't re-encrypt, that works. >> Yeah, I think so. And then there's But also in general, you could just have a Yeah, like the the chain itself like functions as its own like I have set this thing to Bob because Bob can obviously access the chain over some twoe period or something, right? For example, >> that could also work if but if that's we're assuming aness or the part of the goal is we want to have no online assumption of Bob, >> right?

>> Want Bob to be like he thought things working a year later has to go. >> I see. I see. He's really about the long term this year. That's right. Yeah. Okay cool. >> Oh, there's one other observation I forgot to make where I've been saying this in terms of Alice pays Bob, but if Bob pays Alice the first time, like the first them is from Bob to Alice, Bob can always find that really easily. So, we don't have to solve that problem. And you figure that out with nullifiers >> because Bob can figure out when did he spend his nullifier.

>> So, therefore, Bob can always find transactions where he's the first payment. >> Okay. So all of our problem is just figuring out when the first time you receive money from someone >> and this is the biggest case this is a problem is like hey here's my address on Twitter please donate to me >> that's kind of the use case that's really hard which this relayer or sync local sync server solves >> interesting >> cuz everything else is kind of easy cuz it's not a big zod the app already gives a custom address for every payment >> zle being the new Zashi the new wallet >> is Is that already live? Have they launched yet? Okay. Cuz I knew they had like done the company, but I didn't know if the product was already out.

>> Your app should just rebrand now. Like there should be some >> Sashi should be Oh, wow. >> Yeah, this should be live. It happened to my phone a few days ago. >> Okay. Crazy. In terms of the work that you're doing, like you're now contributing to it, but is your company like are you working on a specific part or project? >> So, I'm basically working on what I whatever I think is like the biggest to make the UX of this all win. So there's there's a shielded sync component which is really big. There's a second and there's two other kind of big things. The second big thing is the UX of Zcash. Zcash is a proof of work blockchain with 75 second block time. So I've basically been making a proposal to first drop this 25 seconds and then we want to be doing fallen work to like one keep

dropping block time but over time figure out some way to make a pre-MP like thing from the miners cuz a lot of UX just needs like a good one comp guarantee or even like something zero comp that's better that has like for instance we have 80% of hash power saying hey I promise I'll include this no slashing if I don't That's good enough for most $10 payments or like uh and then you can start getting really good UXs. Right now the problem for Mastercard, Mastercard has like a rule where if in testing you take more than I think it's like 1 1.5 seconds to like respond, you don't get accepted as a credit card.

>> Sure. Cuz your network has to be fast. >> So a 25second block time card already blocks you being a credit card. It's too slow. So, but these pre-MPS could now start doing that for like small even like mediumsiz payments like and now we can get kind of the UX we need the whole meme of you know you don't want to wait an hour for coffee paying for coffee okay let's go solve that let's go make like there's a bunch of different solutions you got to go discuss a very bunch of trade-offs like decide which one and but then make that happen and the third thing is getting Zcash postquantum so what's kind of cool is you can first be postquantum recoverable which means that if a quantum computer came tomorrow, we could just freeze the shielded pools and you could be able to

recover your funds. You could do a we could make a new Stark >> that correctly proves you own these funds. >> So the update you need to make that work in Zcash is tiny. It's just a super small wallet update. >> So where we are right now, even if we drop block times like 10 seconds or even 25 seconds, it's too slow to be a credit card. So, we'll have to do some sort of preconfirmation tricks or finality gadget. There's a ton of ways you can do this. You know, you could imagine there's some committee of uh miners based on hash power, like the trailing hash power of the last 6 hours. And if twothirds of these miners say, "Hey, we'll include this transaction in the next block." Then you as a user are pretty convinced you will be included in the next block. So, there's a bunch of

ideas of this form or you could be inspired by Salana Shreds. There's no consensus on how do there's a bunch of ideas. So after we drop the block time, it'll be time to discuss those to figure out how do you get much better UX for users. >> And then the third big problem we have as well as all of crypto is how do you get postquantum? So to be postquantum has two parts. We can very easily get your funds to be postquantum recoverable. Meaning that quantum computer exists tomorrow. We can freeze the pool and then you can still safely recover your funds >> and them in the postquantum shielded pool >> also without revealing your balance. So that's quantum recoverability. It's a really tiny wallet update. It's actually this should happen pretty soon. And so

that long as you touch your funds after the wallet update, >> it's safe. The quantum computer cannot make you lose your money. The second thing is making a new shielded pool that's postquantum. That one we actually have more time for thanks to quantum recoverability. And so over the next two years, I'm sure we get fully postquantum shielded pool. First we get all the recursive proofs working in the easier mode with the curves. Yeah. And we tried the high scale this year. Like the dream is we get to 500 TPS shielded sync fully killed and like much faster UX for actually writing transactions.

>> Going back to like you originally were a ZK ZK researcher. Do you think is ZK solved? Like now everything we've talked about with you is is engineering. It's like implementing these things and the things around the ZK systems. >> Do you feel like is it is it finished or do you think there'll be some breakthroughs? Hm. I don't think the ZKVMs have gotten it right yet. >> Uh where there's a huge amount of like work that went into the modern CPU chip like how do you make the current the modern CPU so fast and this involves a lot of like instruction set architecture SIMD etc. And I feel like this is a huge amount of complexity in design. Maybe if

you wear your theory hat oh it's all just constants. If you wear like the art of like making CPU chips, no, this is like 40 years of work if not more. The Apple M1 chip was a huge leap over what existed before. >> And you feel like ZK like ZKVMs need to achieve something similar in that way. >> Yeah, I don't think a similar level of like CPU type optimization, how do we do these bulk operations really quickly has happened yet or it's obvious and people care about and work on it, but yeah, I still think there's a ways to go. I'm actually curious your take on this gear.

I feel like you're also even closer to most of ZKVM teams than I am. >> No, I mean I agree. I think like ZKVMs are I I I mean I have a lot of thoughts about the exact mechanics of ZKVMs, how we do them today, why they work, why they don't work, whether we have even adapted them to the systems that we're proving. It was very weird that we spent a bunch of time compiling down to some random language and then use that to like compile back down to like something else and then only after that do we compile to like some like weird thing which then compiles to a PCS. And I'm like what the [ __ ] are we doing here, right? It's like >> six layers of abstraction for the sake of abstraction essentially. And like I get it because you have to balance out

doing it quickly using developer tooling and all this stuff, but it's like come on. Like do we really expect like in the long arc of history we're going to have the six layers of abstraction each of which are just not actually contributing that much and eating through this entire stack? Like the answer is no. >> We have very little PMF of ZK outside of blockchains, especially this recursive stuff that's high speed. >> This client stuff is a very different problem for privacy. And then for all these onchain environments, you're constrained, but with the blockchain support and like there's not been much demand for here is very long compute that I'm going to need to repeatedly run. I'm going to be willing to pay for a six-month even research project to optimize that. The closest real password of blockchain that I can think of is the

Google age verification thing. they spun up internal team to like, hey, let me scan your ID card and give a ZKP that you're above 18 or whatever from a real passport. >> Mhm. >> And that's like a very small client side proof. It's not the same ballpark of ZKVMs. >> Yeah, I agree. I think that there's going to be a big there's still to be a big question about really really big competition. I mean, there's always the joke of like, you know, Intel would make bigger and better chips and it's not because like there was an obvious immediate need for bigger and better chips, just like some someone would figure out something to do with the like bigger and better stuff. Um but I agree right now the like main reason that people are not using or are using ZK that is not blockchain related is

essentially these privacy preserving like ID like technologies which obviously I have some thoughts about some good some bad but yeah >> so okay anyways given all of this I still have the big question Dave for you which is let's say you could wave the magic wand and just like have the stuff for this year next year even even a year after done and it could be it could be whatever you Yeah. What would Zcash or privacy enabled technologies or what does it what does it look like like what's the like you know metallic style manifesto level vision that you would expect uh happen here. Okay. So first we focus on Zcash maybe the manifesto Zcash is in the hands of like billions of people on the planet where we're using it for everyday like private payments.

So this means thousands of TPS on the chain. No one's mentioning this like quantumware anymore. We're just done with it. It's all solved and then it's all fast. Like same way no one talks about what's the clearing time of a credit card. No, no one thinks about the block time of Zcash. This is like not a property we have with like even Athereum main today. We're hit on the 12 12 seconds repeatedly. We need to fix that. It's like all of your normal actions super quick. And so that's Zcash the chain. And then maybe more broadly ar society I want to be building the world where you own kind of all your digital data and you always do proofs about your data besides like you know what's definition of sharing it. I even dream of a world where there's a Twitter fork or Twitter competitor where the length

of time your eyeballs spend on the on a given tweet is not Twitter's data. It's your data that talks to your local LM and your local LM queries some database to say like here's the rough shape of content I want. And so I dream of this world where you have all of your data locally owned and all the sort of global interactions gets privacy. Even more concrete than that that that's we probably are actually going to do is there's all sorts of ways if you do subscriptions like hey I need to pay for an RPC node etc. And importantly, an important part of subscriptions for any API service is that there's rate limits.

We're going to get all of these need to be private of okay, I'm going to provision or pay for a thousand queries per day from I don't call it ZK and Fura or ZK alchemy. Okay, I can do every single query from an anonymous IP. So, it can't be correlated. and attached is some data that's privacy preserving that lets ZK Alchemy know that I'm I'm obeying the rate limits, >> but they have no idea which user I am. So, we're going to need like ZK subscriptions for like SAS stuff all across stack. We have private P2P anonymity solutions and you own all your own data, all your subscriptions fully private. The the merchants whatever can like get their security properties they need usually, you know, they need rate

limits to provide service but not know who you are. I think it's so it's so important and so timely. I don't know if like you're hearing, but all of these companies that people have been relying on for years and years, their communication rails in certain cases >> are making decisions >> post you signing up for them about what they're going to do with your data. And I don't know, like somebody recently shared that like Instagram their chat apparently had been encrypted. It's going to be unencrypted soon. And it's like I I don't use the Instagram chat much so I personally don't care but I know other people do like they keep in touch with friends there and that's not a choice they get to make and the reason is like for AI training and you're just

like [ __ ] you guys excuse my language but like all of those people people I know are just going to be fed into this machine and they did not >> like they were already signing up maybe to be like advertised to maybe they had agreed implicitly to that like, "Okay, I'm going to get shitty ads. That's the free software model." But they hadn't signed up to be like eaten by the AI machine. And I think that's really sick. I think we're we're too late to be honest, but I I hope we can build stuff soon. I mean, also, I mean, I I think I've mentioned this on the show before, but like medical data, like how is it that these industries that pride themselves on privacy are like so

incredibly unsafe and all of your medical data should be in your own possession? Like, why is it sitting on servers in different places? It doesn't make any sense. >> That's my end rant. >> Dave, thanks for coming on the show, man. catching up with us, sharing the story of Osmosis, sharing the story of sort of your return to privacy. I am so happy to see you in our channels being more active again because I feel like you were more off in cosmos land for a while. So yeah, welcome back into the fold or I guess you've you've always been here. I shouldn't say that >> way more now than before. This is the >> Okay, good.

>> I truly believe this is the most important problem to work on. >> Very good. >> Very very cool. All right, I want to say thank you to the podcast team, Rachel, Henrik, Hector, and Tanya, and to our listeners. Thanks for listening.