Umar Roy, co-founder and CEO of Succinct. Welcome. thanks for having me. Before we get into some questions, I want to share a real quick story that I hope will kind of set the stage. A couple of years ago, my brother, my wife and I went to his wedding and my wife was really, really nervous because he's a doctor and all the guests were like doctors and lawyers. She's like really highly credentialed people. My wife is, you know, she's very accomplished on her own, but she was still like nervous about, you know, shoulders with all of these doctors and lawyers. When we got there, I went to get her a drink and when I came back, there was like a crowd around her. And I like peeked in and then tried to like, what's going on? And she's telling everyone a story about her encounter with a shark. And everyone was just like, so enthralled by this. And when I think, when I look at some of the branding assets of Succinct, I kind of feel that way too, in that it's like disarming. You know, here she was initially just very kind of nervous because of these highly credentialed people. When she got there, it's like everyone was just drawn into her. And what I love about Succinct is I get the same feeling. It's like highly technical, super capable, great team, but the branding assets are like really disarming. And so that's really kind of what led me to think I need to get Uma on the show. Like we got to talk about that because it's such a dichotomy. You know, on the one hand, you've got like really credentialed team, really accomplished, but then all the branding stuff is just like, wow, it's like, it draws you in. So I wanted to start out with that as a, to kind of set the stage because it's really, really impressive and it's actually really, really hard to do. And so what I was hoping we'd talk about during the show, I want to go over product, developer marketing, branding and growth, and then, And with the test net because I am ranked number four right now on the test net with 11,833 stars. So I was three yesterday and then some jerk beat me to it. And so I'm number four now. anyway, tell us how you got into crypto. I know you've worked at a number of really well-known companies, but crypto specifically, like what brought you in? First of all, yeah, thank you so much for all the really kind words about our branding and marketing. I do think one of our kind of core company, you know, missions is to make ZK down to earth and accessible to everyone. I think the technology is obviously amazing. It has so much potential, but for so long, it's been so complicated and understandable by, you know, less than a thousand people in the world. And one of our biggest goals for this year and also going forward is really to make it much more understandable to like the common person and kind of see, make people realize, you know, how important it is to like all of blockchains and potentially beyond. So it's awesome that you really resonated with it because that's what we've been aiming for. In terms of how I got into crypto, I've always really been super into math and that's kind of been a common theme throughout. high school, college, and even afterwards. And I was doing stuff in AI. So I was doing NLP or LLM research at Google Brain actually before GPD 3 came out. So it was kind of like the older days. And so I was doing that. But then on the side, I've always kind of liked crypto's principles and values of just this kind of exciting wild west, like underexplored. libertarian, know, very freedom oriented, permissionless systems. I've kind of always liked those values a lot. And I started learning about ZK in particular, on the side when I was working, just because I thought the math was really interesting. And that was actually like my original hook in. And then I just started reading more and more on the internet, getting deeper and deeper. And then eventually that led to starting to say, That's cool. My first encounter with ZK was when I interviewed for a job with Oasis Labs and Dawn Song was one of the people that I eventually reported her. She was like my boss and she's like a big deal in ZK, I didn't know anything about ZK. I studied math in undergrad and then for grad school I studied computational linguistics. So got into kind of the, before AI became like a big deal, like this was like, before LLMs of course. We were dealing with like small corpuses or corpora at the time. And so that was like really intimidating, but she was able to help me kind of better understand it in a way that I was able to then help developers because even with developers, ZK is such a new thing. And what I learned, at least when I led developer marketing at Oasis is that a lot of developers kind of, they don't necessarily... care so much about the technology, they just kind of want to build something. And they just kind of want things to just work. And so that was kind of became my focus is like, how do we focus on like education, but really help the developers kind of build stuff? Tell us about the product. There's a lot of ZK VMs on the market, maybe not a lot, but there's a few, right? Tell us kind of how Sysync or SP1 specifically kind of stands out maybe in relation to a couple of competitors that you view as competitors. Yeah, so what we're building at Sysink is a ZKVM called SP1. And what a ZKVM is, in very simple English, is it lets you use ZK by writing normal code and taking normal software and proving it. So more concretely, you can just write your program that you want to be verifiable and generate a proof for in Rust. And you can just take that Rust code, stick it in SP1, and then you get a proof that you can use on chain. for things like roll-ups, bridges, or co-processors, really any application where you want to do a lot of computation on chain for very cheap. And then we're also building a prover network, which is kind of this protocol around our ZKVM, which you can kind of think as kind of a cloud offering, but decentralized for SP1. So SP1 is a piece of open source software. You can run it on your laptop, but... The reality is that it's not super performant on normal computers and you really need GPUs and specialized hardware to efficiently generate proofs with it. So our prover network is this two-sided protocol where you can send proof requests with one API request. And then on the other side, there's this distributed, decentralized set of provers who are running data centers with GPUs, generating proofs, and then sending them back. And the best part is because in ZK, the whole point is the proof is self-verifying. Like if you have a proof, you know that the computation was done correctly, because that's the whole point of the proof. Anyone in the world can generate a proof and get paid for it. And it doesn't matter who they are, where they are. You don't have to trust them. Like the proof is the proof. So yeah, those are the two things we're building. There's like the ZKVM SP1, and then surrounding it is the decentralized prover network. Gotcha. Yeah, I'd love to get into the decentralized prover network because I love two-sided markets and that is clearly a two-sided market and I'm curious to know how you hope to build that out. I imagine that's going to be after Mainnet or maybe in conjunction with Mainnet. But let's talk more a little bit about the product. As you're building out the product, you're now in test net phase one, which is a crisis of trust. I love the storytelling, by the way. So let's talk about that. I guess, what are you hoping to test in crisis of trust in level one and some of the metrics that tell you whether or not things are going well or maybe you need to improve on certain things? Yeah, so we're building out our prover network. And right now it's in test net phase one called crisis of trust. And the theme of the test net is basically this video game style test net where there's a crisis of trust where you can't like trust what media you see on the internet. You can't trust like. images and videos and like video game scores and all that kind of stuff. And the theme is basically this dystopian future where, there's no trust on the internet and then ZK and verifiability can help save us. So that's like the setup and theme and story. And the test net in stage one is really meant to open it up to one side of the marketplace, which is the people generating proofs. So basically with this test net, you can play these like games on our test net. And then many of the games have a proof generation component where once you're done playing the game, you generate a proof and you actually use, you're using our prover network to generate these proofs and get rewarded for testing out the software and the network. So for example, there's one game called like Ethereum, the last block where you select a random Ethereum block, you generate a proof of the Ethereum block. And you know, the theme of the video game is like you're helping Ethereum scale by generating these proofs. Another game, I think this one's pretty popular is ZK Flappy Bird, which actually we recently had to rename due to some like, you know, think trademark or whatever considerations to floppy GPU, but a similar theme where like you play this video game and then there's actually a proof of kind of all the actions you took. to get to your high score. So you generate a proof for that and then yeah, you get like rewarded based on like your high score. So I think the theme of the test net, which is really awesome is it shows you can use ZK, especially with our tech on like things like blockchain use cases like Ethereum, proving normal pieces of software like video games and a bunch of other stuff. So yeah, that's a theme that ties the test net together. That's cool. By the way, the last block is like the worst game because I think I played it twice and then like all my 10 USDC was gone. It was like, what? But thankfully, yeah, it's like actually really hard. Like you don't do anything other than like you just lose your USDC, but Flappy Bird is also really hard. Like I think the most I, or Flappy GPU, I think the most I got was like past like the third pillar. And then it just like. I would die, but you know what's really fun is the racing game Turbo. That's pretty fun. Oh yeah, that's a more educational game that kind of like our new ZKV, one of our distinguishing factors, I think you asked me earlier is like what sets us apart from our competitors. One of our distinguishing factors for our ZKVM SP1 is that it's really, really fast. And our latest release, we're always updating it and improving it and making it better. We called it SP1 Turbo. And yeah, SP1 Turbo is super fast. It's like the fastest ZKVM. It's up to an order of magnitude faster on a lot of use cases that our customers care about, like blockchain use cases. And then a dev on our team built this racing game kind of inspired by how fast it was where like, you you have like a turbo bike and then you're collecting, you know, chips and it speeds up and it is a very fun game. I was actually addicted to it in the office. Like I sometimes play between meetings, which is bad for productivity. But you're proving, so you're helping with transactions. That's good. Yeah, I mean a lot of the games do generate proofs and so it's like helping battle test our software and we actually did with our test and we actually did Find some things in our test net that helped us from all the traffic and load helped us, you know improve our systems Which is the whole point of a test net, which is awesome let's talk about that. So all of these games, generate proofs. So that helps with generating transactions and proofs on the system. The prover network, who are currently the provers on the network, since it's not entirely decentralized yet because you're not a main net? Is that you guys? OK. Yeah. So right now in phase one, it's like anyone can request a proof, but we're the only approvers. And then in phase two of the test net, we're going to onboard a bunch of approvers and make that open. anyone in the world, basically, or not anyone in the world, but we'll start having approvers other than us. So that's phase two. And then main net will be like, it'll, it'll be fully open system with like our auction mechanism for pricing and stuff like that. And so for those that are listening, this might be the, so phase two might be thought of as like getting validators on the system, other than, except validators, these will be approvers. And so you'll have requirements on like what it takes to become a prover, I'm guessing. Okay. Do you have any idea what kind of machine you'll need, anything like that yet or no? I think we'll have a few options for people. basically, I think one of the options will be participating with a GPU or your hardware at home. But then we also want to make it an inclusive experience, because obviously it's not like everyone has a spare GPU lying around. And so there will be options for people to participate in other ways. But yeah, you'll see this all in the next few weeks. Yeah. Do you envision like large data center, like institution type kind of provers also on the network? that's, I see kind of as projects go to main net, you know, they start out with kind of a small set of really institutional validators and then they open it up to kind of everyone else. Do you see something similar? Yeah, I think that will probably be a likely path as we kind of try out our test net phase two and then main net. But yeah, I think a lot of that will also be influenced by how the phase two of the test net goes. Yeah. Okay, so we've talked about how, so phase one, it's generating proofs on the network. then let's talk about like, as you think more about having developers, actually, let's talk about target customer, like who's the target, initial target customer for SP1. I imagine it's chains primarily, rollups, L2s, maybe applications too. Yeah, so some of our biggest customers are rollups today. Basically, I have this thesis that in the future, every chain is going to be a ZK chain, where basically you post your data to an underlying data availability layer that's doing ordering and consensus. And then you use ZK to scale execution. So that layer can be Bitcoin, can be Ethereum, can be Celestia, or even something like Solana. They're kind of starting to flirt with the idea of these ZK scaling with their network extensions, which are really roll ups. So in the future, my thesis is just almost all transactions will be ZK proven using a ZKVM like SP1. So yeah, the biggest PMF we have today is definitely on the roll up and chain side. But we also do have a lot of other applications that are building. So for example, bridging is a big one. Another one is like replacing trusted oracles with trustless oracles. So for example, Lido, they're prototyping this ZK co-processor where instead of having a trusted oracle that dictates everyone's stakes, you actually generate a proof of like what everyone's staking and rewards are. And there's a few other. you know, interesting examples of things people are building. So another one at a hackathon recently that I thought was really cool was a tested image editing where basically you can take an image and then you can prove that it was transformed in a certain way, like cropped, blurred, whatever. And then you can show that it originated from like a real camera as opposed to like from mid journey or something like that. So I think That's also really cool use case. So yeah, I think one of the awesome parts about ZK, especially with SP1 is that because you can prove normal code and normal software, it's a dramatic expansion of where you can use this verifiability. so, yeah, rollups are the biggest things today, but I'm also really excited about all the other software that's gonna get proven. Yeah, let's talk about the roll ups. I know that there's a point of view where roll ups are going to be everywhere. It's going to be kind of like a thousand flowers blooming is kind of the idea. Like philosophically, initially anyway, a lot of people thought that was a good thing. And then as more and more roll ups kind of went into the market, very few are actually differentiated from each other. And so we're kind of learning more that, you know, I don't, maybe it's not the best thing to have, you know, make it so easy to, use a rollup as a service provider, just spin up your, your, you a rollup. What's your point of view on that? Yeah, I think generalized block space only really makes sense for people who already have distribution. And then you want to build an ecosystem on top of that distribution. So for base makes a lot of sense. They have a lot of distribution through Coinbase and they're a big enough platform where having base as a general purpose platform totally makes sense. You know, maybe I think there's other projects with that level of distribution, a few other projects that could also have that story. But then I think If you're just a developer and you want to build an application, a roll-up still has a lot of benefits, right? Like you get your own dedicated block space. You can do more customizations. You don't have to rely on another general platform. like, because, know, who knows what the best general platform will be in like two, three years. You can kind of like have your own sovereignty. So I'm actually a really big fan of like application specific rollups, where basically it's just like the rollup is the new server and then you just build your app on it. And I think the pitfall of the RAS stuff in a way was that maybe a lot of people were just trying to launch these general purpose platforms when they really didn't have the distribution or the reason to do so. But I'm much more bullish, like an application team being really targeted and being like, Hey, we want to build our application on a rollup. and like, let's launch it. And I think we will see that theme coming more and more. And so you're kind of describing kind of the hyper liquid approach. I wonder, are there other teams like that where it was application first and then the chain kind of was almost like played second fiddle almost. Yeah, think Hyperlink is a great example where they built an application specific chain and now they take that distribution and mind share and like liquidity and excitement and users and then they're building a platform on top. And I think that's absolutely the right approach. I mean, arguably Coinbase and base is kind of an example of this. It's not as direct, right? Like Coinbase bootstrapped the platform and the attention and users through their centralized exchange and now they're building base as their general purpose platform. You also see this with BNB chain, right? Like same pattern. Yeah, exactly. Yeah. Exactly. Unichain is another great example. World even, Worldcoin, right? You have the orb, you have this very specific application and then now they're building their own general purpose developer platform, but it ties back into their original application, right? Like it ties back into kind of like, have human prioritized block space with the world coin IDs, all the applications integrated into their world ID system, etc. In terms of distribution for SP1, I imagine you'll probably take advantage of some of the roll up as a service providers and SP1 becomes part of kind of kind of selection process of like, which data availability do you want, et cetera. Is that kind of, am I thinking about that right? Or are you thinking of other ways to distribute SP1 across developer kind of chains? No, I think that's exactly right. Actually, we work with lot of RASs pretty closely today to kind of offer our integration into OP stack as a like a ZK stack. So we work especially a lot with Conduit and they deploy their roll up with Conduit and then they also use OP succinct, which is our offering to take any OP stack roll up and make it a ZK roll up through Conduit. And then Conduit just sends our network a bunch of proofs. Yeah, Conduit's great. I've worked with Conduit, also Altlayer, and I'm actually interviewing someone from Gelato in a couple of weeks. So that's really cool. When they use OP Sustained, does the rollup become part of the super chain in that case, or no? Sorry, give me one sec. I'm gonna see if someone vacuuming in my office. It's like the noises. yeah, yeah. yeah, so you were asking about the roll up. Yeah, yeah. if I work with a RAS like Conduit and then I choose OP Sysync, does my rollup then become part of the OP Superchain or how does that work, I guess? Yeah, so the super chain is the set of chains. Well, the super chain hasn't launched yet, so it's hard to say exactly what it is. But it's a set of unified chains that all can interoperate with each other. If you use Opie, so to be part of the super chain, you actually have to like do a bunch of things like you have to do sequence or rep share with optimism. You have to follow their fraud proof game. You have to use Ethereum DA and Ethereum as your gas token and all this other stuff. So you can use the OP stack, but not be part of the super chain. And that's actually pretty common. So for us, if you use OP Sysync, you actually don't become part of the super chain because you're using your own like proof system, basically. So yeah, most, a lot of the rollups, they're just like, I want to use OP stack. I want to be fully ZK because I don't want the seven day withdrawal nonsense. let's use OPCSync and it's very easy to deploy and it's relatively inexpensive. So yeah, that's like our main kind of customer base. Yeah. So we've established that the roll-up as a service providers are one of the kind key distribution channels for SP1. What other ways are you trying to reach or how are you reaching developers and chains to learn more about SP1 and work with you? Yeah. So RAS providers are a great partner channel for us because they know all the roll up teams, all the chain teams and like all the people who want chains. So we work really closely with a bunch of RASs including Conduit. And that's been a super productive, I think positive some relationship. And for them it's awesome because a lot of their customers want ZK. They don't want to deal with the seven day withdrawal stuff of all the optimistic sacks. And so they can tell their customers, Hey, we have a solution for you, which is OPCsync. In terms of other use cases that go beyond rollups, mean, a lot of developers just know about us through like Twitter and like, you know, public presence. We run hackathons, which is also, I think a good way to get the next generation of new teams working with us. We hosted this in-person ZK residency for a month where people applied and came to cowork out of our office. And we're thinking of hosting some things some more in the future. So. Yeah, those are kind of like all the activations that it's a broad strategy. I think you have to try everything to attract developers, but those are some of things we've That's really cool. Let's kind of go into the branding side. Succinct has taken advantage of, not taken advantage of, but has really utilized some pretty amazing videos using real people. And I think that kind of, it's quite different. I haven't seen that done very much within the crypto space. What led to kind of that kind of approach. I mean, that's a lot of our head of marketing's genius. So I really can't take credit for that. And he deserves all the credit for it. I think at the highest level though, the philosophy always was we don't want to do what other teams are doing. Like we don't want to do the Logo X Logo partnerships. We don't want to just do the same kind of 3D visual marketing videos that with like the blocks, you know, going into space. that everyone else is doing, we want to do something different. And so, yeah, and our head of marketing is also a really big fan of that philosophy and a big creative genius. yeah, I think there's always this philosophy of breaking the meta and being different and standing out. And then also making ZK much more relatable to the average person and not this abstract, futuristic thing, but more just like any other advertisement. or commercial for like any other real product. So I think those are some of the inspirations. But yeah, I really cannot take credit for it. think a lot of it's due to our amazing head of marketing and also the creative genius of a lot of other people on the team. Well, I love the word relatable because it really is. You know, when you see other people that kind of look like you, it may not, and it's not like specific to like the ZK kind of, you know, meta or SP1 as a, as a, as an offering, but it's like relatable. It's memorable. And like we're talking about it right now. And it's actually kind of one of the first things that drew me into this, into like me wanting to learn more about Succinct. And as I learned more, I'm like, wait, ZK? What's going on? Like it was actually kind of pleasantly surprising. you kind of want more experiences like that because it's pleasantly surprising and it immediately becomes memorable. And I may not be the target customer, but now I'm like a fan. And so it has that kind of benefit of using real people, the branding at the... the branding kind of aesthetic, you has created a fan that's not like a target customer, but now I'm like, I'm going to promote Sysync because of that. So that's really cool. Going into the test net, like I'm curious about, you could have approached the test net in lots of different ways, right? You kind of, you guys arrived at this like really kind of fun approach that is very inclusive, that includes that is welcoming to non-technical people, technical people, just really anyone who wants to play around with the test net. But you could have taken it a very kind of different way, where it's kind of stuffy and boring and very kind of creating circuits. Why didn't you go that way? What was the thinking behind the test net? And I guess, what were some of the internal dialogues, dialogue that you guys had as a team? Yeah, that's a good question. For us, I think when we came up with the concept of a test net, we really view one of the biggest problems with ZK as it isn't understandable or relatable or accessible to the average person. And for our developers, the magic of our product is that now any developer can use ZK, right? Like with what we built with SP1, it went from maybe a thousand people in the world, but realistically, honestly, more like a hundred people being able to use ZK to being like millions and millions of Rust developers or really any developer. So it was like that moment, even in our very technical product of being going from like hundreds to millions. And with our test net, we were like, well, do we already want to target the people that are already excited by ZK because they already know about us. So it's not that interesting. we wanted to really take it mainstream and like break out of the ZK bubble. And so internally, when we were coming up with concepts for the test net, we were always like, what would be exciting for like one of our normal friends to try out and play with and make it relatable and understandable to them? And our head of marketing, he's very like non-technical. He was actually a really big source of inspiration because he kept on saying, I want to ZK video game. I want to prove a video game. That would be cool. And I think that was kind of the initial seed for like the whole branding, the whole story, the whole inspiration and how it came together was really born out of a desire to make it engaging and fun for normal people. But yeah, that was always our top level goal when designing it. And everyone was a line. No one was no one argued. No, I want it like stuffy and technical and like hyper technical and boring any arguments that happened internally. No, no, I think we were all, I think we all on our team were like, we all were like, the number one problem with ZK is like, it is really stuffy and boring and technical, so we need to fix that. So there was always alignment there. That's really cool. How do you explain ZK to your mom? I'm assuming your mom is not technical, but maybe she is. Yeah, I mean how I'd explain ZK to my parents. I would say fundamentally ZK offers you this primitive of verifiable computation. So you can prove to someone that something happened or some computation results in this answer without revealing to them what all like without exposing everything about the computation. And then also so that they don't have to redo the computation themselves. And that's maybe like the easiest one to two sentence explanation. When I led developer marketing at Oasis, we had a big campaign where we invited the community to create ZK jokes and then submit them. This was before chat GPT. So people actually had to be creative and think. And so I hadn't used chat GPT or Grok for this, but I actually asked it, hey, tell me a ZK joke. And it told me one. And so I want to see if you think it's funny. So why did the blockchain developer bring SP1 to the party? Because it could prove it had fun without revealing how it happened. This is kind of funny. It's not bad for a chat GPT joke about an esoteric topic. When we did this campaign at Oasis, had, I don't know, maybe less than 100 submissions from the community. But what was really cool is that they were forced to really think and study what ZK was and then actually create a joke out of it. What's unfortunate now is with ChadGPT and Grok and Claude is they could just use that. And you almost bypass thinking. which is unfortunate. That was actually a campaign that went pretty well. But yeah, that was just a few years ago and it's so crazy that like, chat GPT wasn't even around then, you know, but now it's like so ubiquitous. Okay, so I've established that I'm ranked number four on test net with 11,833 stars, which has been really fun by the way. I have a question around like, So I poked around on the desktop and in the trash bin is a picture of Nick White with actually three, it's like one picture, but there's like three Nick White, like his face is on like three bodies. What's going on there? desktop has a lot of cultural references, I would say. Nick White is actually a great friend of the company and me personally. So we're, we're always making memes of him that it's, kind of, I don't know, it's kind of like a inside joke almost. So yeah, that's how he ended up there. That's funny. I know Nick really well. He and I worked at Harmony together. got it. Yeah, yeah. Nick's awesome. We love Nick. I think Max Reznik's also in the trash bin. He's also actually a good friend of the company. We were just making fun of him for leaving for Solana. Yeah, that is so cool. And that's the other fun thing about the test net is that there's like really fun kind of it encourages exploration. And the exploration may not like, you know, lead to point to stars, but it's like, it's fun that we're talking about or even right now creates conversations, you know, and so it's what can you tell us about phase two with prove with us? Any any, I guess? I don't know, any hints that you could share with what might happen then. Yeah. So as I mentioned earlier with phase two, we want to open up the other side of this two-sided kind of network and marketplace. And we want to be able to have anyone prove and, or not anyone, but have people be able to participate in proving. So, so far the test net people participate in requesting proofs. And the second thing we're going to have people participate in generating proofs. So. Yeah, there's, and it'll be pretty inclusive. So you don't need to have, you don't necessarily need to have your own GPU or hardware at your house. There'll be options kind of for everyone. We want to make it maximally inclusive and kind of a fun experience for people to be on the other side. That's so cool. Yeah, I'm excited for that. I mean, now that I'm in like, I've participated in phase one, like phase two, and I'm curious to see what will happen. And I'd love to participate. I might have to get my own GPU though. No, no, you don't have to get your own GPU. But I want stars, so I've got to get competitive that way. No, there will be options for everyone. yeah, don't, no one should go out and like rush by a GP or anything like that. Yeah, it'll, there'll really be like, people can participate no matter what and people will see you soon. Awesome. Well, Uma, thank you so much for taking the time to speak with us. You guys are killing it. And I think are a good example of how to make a highly technical product like friendly to normal people. And in normal people, I mean, not just like non-technical people, but a lot of developers fall in that camp. know, most developers just kind of want to build something and not have to worry about the kind of the plumbing underneath. And so being developer friendly. is I think you guys are well on your way to accomplishing that. so kudos to you and the team. And thanks for taking the time to spend with us. Yeah, for sure. Thanks for having me on and yeah, thanks for enjoying out our test net, enjoying it. yeah. Thank you.