Amrit Kumar from AltLayer, welcome. Thank you very much for inviting me. Been excited to do this pod with you. I've known you for a while and have watched from the sidelines all the cool things that you have been up to, starting with Zilliqa and all the research into interoperability before that in sharding. You're now at AltLayer. If you were to describe what AltLayer is to, let's say, your grandma or your mom, how would you describe what you do and what AltLayer is? So we are technically a tech service provider. So people come to us and they want to have let's say infrastructure, a blockchain essentially, and we help them get that. That's basically in a very nutshell, this is what we do. Yeah. And do think that's sufficient enough for your loved ones? Okay. That's cool. By the way, what's a blockchain? Do they ask additional questions like that? I mean, if you go there, it doesn't end. I think it becomes a rabbit hole. But yeah, basically, I think if I explain to my loved ones and say, I'm helping people deploy, know, crypto infrastructures, they kind of get it. And that makes sense. And I imagine with the holidays coming up, you'll probably get a lot of questions like that. Like, what do you... I can't explain Eigenlayer and AVSs. That's a bit more complicated, which by the way we're also involved in, but explaining simple blockchains. People know Bitcoin, people know Ethereum. So worst case that I have to say that, imagine Bitcoin, but built just for you, just for your business. yeah. Well, since you mentioned Eigenlayer, AltLayer recently received a grant from Eigenlayer. You want to tell us what that's all about? Yeah, sure. Of course, we have been deeply engaged with the Eigenlayer ecosystem, I say from the very start, even before Eigenlayer's core contracts were publicly available. And so we have played a role, we have done different things. we, for example, we started with an AVS builder. So we built AVSs, so we had a framework called ReStake Rollups. So basically it's a service. service. Got it. it's, so imagine any service that you want to provide to either end user or to businesses, but that service is decentralized. So you have a bunch of nodes, bunch of operators, bunch of service providers offering that specific service is basically what AVS is. And so we have built several AVSs in the EigenList ecosystem. We have been operators for other EVSs in the EigenLay ecosystem. We have been doing co-marketing and a bunch of other things. For example, we have hosted events together with EigenLay ecosystem. So I guess in recognition of all that past work, we got this grant that we decided to pass on, pass it on to our community, which is the community of alt holders and re-staked alt holders. That's great. And that's quite generous of you. But a really, really great way to more recognition for the AltLayer brand as well as users of AltLayer. So that makes a lot of sense. Now speaking to crypto native folks, how would you describe AltLayer? Yeah, so, I think we have expanded our product services, product offerings to more than just rollups. Rollup as a RAS, so basically we help people launch rollups. So imagine, you know, back in 2016, 2017, people are launching their own chains. So for example, there was EOS, was Zilliqa, there was Solana, there was Avalanche and so on. After optimism and Arbitrum, people started to think about roll-ups and basically it's a chain that relies on a certain base chain like Ethereum for security. So you don't need to have your own validators as in any other L1. So you basically rely on Ethereum for that security. And so what actually happened was Optimism and Arbitrum, launched, Optimism first by the way, they launched their stack. So basically it allows anyone to basically launch their own version of Optimism chain, customized for their own use case. So imagine you are building your own gaming application and you would like to have something similar to Optimism, but you want to customize certain components. For example, you want to short a block time, you want to increase block size or so on. that you can't do with a chain that you can't control. And so a lot of people who are building games, they wanted to own the infrastructure so that they can offer better gaming experience to end users. so Optimism released their stack. Basically, it's a framework that allows you to basically spin up a version of Optimism for your own specific use case. And so that was launched, which actually encouraged a lot of people to build their own chains, because this was a much easier way to launch chains compared to using something like Cosmos SDK, where you need hundreds of valid reactors and so on. So it was quite troublesome to actually spin up and bootstrap a chain like that. But rollups are much easier because you could basically need a centralized system or sequence, and that's it. And so still, there were a lot of complications in the sense that you didn't know how to set up your sequencer, you didn't know how to change those things. And so the idea for RAS providers like AllClear was to make it easier for you to do that. So you would come to us and you would say, hey, we want a chain and we want these customizations to my chain. Could you help us do that? And we would do that. We would host that chain for you. Not just for hosting, but also we would upgrade chain so whenever Optimism pushes a new upgrade, we would do that for you so that you don't have to worry about that. We would give you any telemetry. So if you want to measure on different things, how your chain is doing. how to, let's say, check whether your RPC endpoints are taking enough load and so on. So that would basically be covered by us. So from a DAP builder perspective, you would focus on building the DAP, building the game, building the DFA application. We would handle all the infra component that needs to be there for that chain to be fully operational. So that's kind of what we do on the RAS side of things. Got it. And RAS is R-A-A-S, roll up as a service, just for those that are unfamiliar with that term. The primary customer for a product like All Air, can you maybe describe what that primary customer looks like and what they're looking for? Right, so about in 2017, or even later in 2020, would say until 2021, if you were a DeFi builder, so if you're building a DeFi application, like let's say, Harvey compound or Uniswap, you could have imagined to build, to building your own chain. People thought that you need to be able to compose, you need to be to work with other applications out there because, you know, for example, let's say you have a, lending protocol like Aave, you are borrowing against some collateral. At some point, you want to connect that to a Uniswap like Dex, for example. And so people always felt like if you're building a DeFi application, you need to be able to sit on a chain where other applications are already built so that you can work together with those applications. But DYDX changed that. So DYDX, was one of the biggest, know, perp decks on Ethereum, not Ethereum, most, yeah, Ethereum then went to Stockway and so on. They decided to build their own chain. And that encouraged other people to consider having their own chain for whatever use case they were exploring. And then with OPStack and other stacks that came out, that made it even easier. So you don't even have to worry about hundreds of ideas. You basically spin up a node and everything kind of works out. And so I would say the initial or the first application builders who went towards building the own chain or coming to a RAS provider like us were mostly game builders. And their motivation was very different from DeFi people. So their motivation was, look, I'm building a game and if I'm building a fully on-chain game, all my transactions will go on-chain. Every single interaction that user is doing or making will be an on-chain transaction. And that would be painful if you were to go and build on Ethereum because Ethereum has a 12 second block time, which means that if you send a transaction, will take about 12 seconds for that transaction to be finalized, which is not really good from our gaming experience. And so they were looking for a chain which they can own and customize to be able to offer better user experience than you would get from building on Ethereum, for example. So that was, say the... Initially, almost every single client that we're speaking with, they're all game builds. So either fully on-chain or Web 2.5 style games or something between Web 2 and Web 3, but most of them were all gaming applications. Some of the people were actually trying to do social stuff as well. So something like Frontek or there were a couple of other games, companies that were kind of exploring how can you build the next Twitter essentially on-chain. And again, the idea was that if you were to capture even 10 % of what Twitter activity is, then you need your infrastructure. It's impossible to build that on any major L1 out there. So that was kind of the reason why we were seeing quite a demand from people who were building games or social, anything that's very consumer specific, very retail specific. But then with DYDX, when DYDX came in and they decided to build their own chain, then we started to see DeFi applications as well, particularly applications. that didn't need too much composability. They didn't need to talk to other or connect with other apps. So for example, Pubdex is a good example, which is literally isolated. So most Pubdexes are sitting on their own. You bring your USDC, you bring your stables, and then you can start trading on them because they're basically synthetics, writing sensors. And so yeah, so I've seen examples on games, social applications, people building taxes. These are the major application scenarios that would need a chain for their own. So it sounds like a transition from developers building their own, building an application and then later deciding, you know, maybe we need this on our own chain because we have different requirements for our users have different requirements than what currently Ethereum layer one can provide maybe something like that and value capture. like, let's say, imagine if you are Uniswap, right? You do billions in volume, know, every week or something, And then... Like from a user perspective, when you're interacting with Uniswap, you're paying two fees. One fee goes to Ethereum validators because you're sending a transaction and that transaction needs to be processed by Ethereum validators. So you're paying one fee to Ethereum validators. The second fee they're paying is to the LP. So people are providing liquidity and you're trading with that liquidity. And so you're paying some fee to whoever provided that liquidity. And some of the application builders felt like paying the fee directly to Ethereum doesn't actually give them anything. And it's also a bad user experience from their perspective. So felt like, can we actually remove that? And if we remove this Ethereum validators directly, then we can actually make more money. So either the DAB builder will make more money, or whatever money they make from this taking away the Ethereum fees, that would go or share. with the users of their chain or their ecosystem. So that was another motivation where they wanted to make sure that they capture more value than what they're capturing today on our general purpose chain where they have shared liquidity, shared iterative base. And that makes sense. currently with Uniswap or the Uni chain, the fees that they collect do not currently go back to the users. It stays with, yeah. But eventually maybe in 2025, they may turn on the fee unlock feature perhaps and distribute it to users. So right now they have this, I think, UI fee where if you, so if you directly use, let's say, the contracts, when you use that contract, you don't pay a fee. But if you are using their UI and their wallet and so on, then you have to pay a UI fee to Uniswap. And then, but there's no switch, know, fee switch. It means that, you know, whatever money comes to them, it doesn't go to uni holders or token holders in uni swap. Yeah, one of the predictions I'm seeing from, you know, on X, know, people come up with, know, their top 100 predictions for 2025. that I'm seeing a lot of is, you know, DeFi applications will turn on the fee sharing module and start sharing the fees with users or token holders. doing it is regulatory and compliance. mean, there's no other reason. It's not that they don't want to share. They're not doing it because they feel like it's risky, compliance perspective. That's the only reason. Yeah, I'm hoping 2025 on at least with the new Trump very pro crypto administration, things will change and that will be quite exciting. You know, on that topic, what things are you looking forward to with the pro crypto stance that the US regulators are now kind of? Of course, now that I mean, by the way, the state of crypto has changed, you know, this year, certainly, with ETFs coming up, people who have been fighting for a spot Bitcoin ETFs for quite a while, and most that they got was futures ETF. And so now you do have spot ETFs. So I also feel like this is just my personal prediction. I feel like the Usual cycles that we're used to, which is like the Bitcoin having cycle, you know, every every few years you will have a very good bull market and then after that you'll have bear market for a long time. I think that might change a little bit. I'm not saying so completely, but I think this this valley will be a little bit more less less deep in that sense because I feel like now because of those ETFs in some sense. you're getting closer to stocks and something even though they're not exactly stock, but the market dynamics will change a little bit in my opinion. So I feel like the previously we were like this when you were going through the bull and bear, now it'd be slightly flatter, I would say, in my opinion. So that's what happened. The other thing that we obviously seeing is a lot of US founders lately have been setting up entities in Cayman Islands and whatnot and they're hiring. I mean, there was someone in Twitter, some joking saying that you can't find lawyers in Cayman because all the lawyers have become directors of foundations operating for US entities. So I think that will hopefully change because, I mean, there's still immense talent in the US and... It's kind of a shame that you have to find ways to bypass or work around rules that don't exist really. You're just speculating that if I do this, I might face issues. I think hopefully that will change. We should also encourage more people to stay in the US and build from there. I've heard stories where people have actually moved out of the US and are living in Cayman, literally living in Cayman. to be safe from all these things, right? Which is crazy. So I think those things will change. And I also hope that, especially from a DeFi perspective, as we were discussing, I think these fee switches and those things will probably become active when people feel like, okay, this normal to allow part of your revenue to be shared with your community holders or token holders. And it's... somewhat different from sharing your dividends. At least if there's some clarity around that, that'd be immensely helpful. And I'm hoping that we'll get that with the new administration. I agree. I think the regulatory clarity will be helpful and help prevent these kind of preemptive workarounds that, know, just out of fear because there's lack of clarity. Now, going back to Altlayer's mission. Now, the promise of these L2s initially from what I remember was, you can build any L2 you want. It'll be very easy to do. And... focus on the specific audience and applications that your L2 is designed for. And so you pick a niche. And then the last year, we've seen so many L2s spring up. With the help of AltLayer as well as other roll-up as a service providers, it's become very easy. Do you think that the initial promise of L2s is coming to fruition or is it still lacking from what we were initially kind of thought. I would say I'm disappointed, to be honest. The reason is... The idea was that, yeah, you will have these OP stacks and can orbit from orbit and you can experiment in all sorts of ways. You can customize in all sorts of ways. Unfortunately, people stopped customizing. So what ended up happening is you have the same vanilla stuff being deployed X number of times or N number of times. And that's pretty boring. There's always Solana guys, people in the Solana committee will always say, look, what did you really achieve by building? 10 different copies of optimism. And I get their point of view. like from a simple switch, I see where, Vitaly wanted to do with this roll up centric roadmap. But I think it came to a point where there's literally no difference. And I can see why people have been pushing for this homogeneity in these systems because, Imagine you have your own chain, you customize in n sorts of ways. It's completely different from how optimism works. Someone else does the same thing and it has its own flavor of optimism. The problem in the system is once you have these different kinds of chains floating around, it becomes very difficult to actually draw a break between them. Because your block size is too big and that won't fit in my block and so on and so forth. It becomes way too complicated to actually interoperate. And that's kind of why it kind of makes sense to actually have homogeneous chains around. Which is fair, but it kind of ruined the whole idea that people would experiment because people aren't experimenting anymore on this infrastructure. People are experimenting on the DAP side of things, which is fine. But then everyone has his own chain. So I would say that I'm a bit disappointed to see there's no real difference between, I don't know, base chain or Unity has some differentiation. They have this, they're this with TE and a bunch of other stuff. So there's some differentiation, but otherwise, and by the way, this is a thing that, a behavior, a phenomenon that I'm seeing, by the way, where people felt like, okay, here's, I'm building my game. I want to customize my thing, but they realized that they should not customize it because it'll make things complicated for interoperability. And so they said, okay, now I'm gonna build my own chain for this. And they built their own chain. They realized that, okay, now I don't have enough users on my chain. And so there's no point in pretending that it's an app chain or app specific chain. have to make it general purpose because I can then invite other people to come and build on my chain to bring more activity because my game is not bringing activity, which means that people are going in the circle where they say, I need an app chain and then go to, okay, I'm... I'm going away from building an app chain to building a general purpose chain and then going back to say, now where do I actually sit? Because now the moment you become a general purpose chain, you're competing with the likes of base. And that is quite difficult to compete with the likes of base. And so there has to be some differentiation, not just at the app level, but also at the info level. Like, do you... Like when you're building, but this is something that I've always thinking from a Web2 perspective, The info that is supporting Twitter is not necessarily the same info that supports Facebook. It's a different, of course the base nodes and the AWS are the same, but in the end, there are differences between, you can't just take whatever Facebook is running and try to build, install Twitter on that. It's not gonna work in the same way. And so I think do feel like there has to be some differentiation in the way people think about their chain, which I guess Cosmos allowed you to do that. A lot of Cosmos chains customize in all sorts of ways, which is sadly not happening in the Ethereum space. It all looks more or less the same. Maybe 1 % of the chains look different, 99 % of the same. So I'm a bit disappointed because of that, to be honest. I agree with you. I've been quite disappointed too, and having been involved in a few layer two projects. To steel man the other side, starting a layer two, even though from an infrastructure perspective is easy, thanks to services and service providers like AltLayer. What ends up happening from what I've seen, the experience that I have is all of these layer twos, they eventually need shared services like an Oracle, like you know, some basic DeFi so that there's some liquidity. And because all of these layer twos essentially need the same types of products and services on them, they end up looking the same and they end up also competing for an Oracle, you know, from the very few Oracle providers that there are. And differentiating becomes very, very difficult. Now, some that are focused on, let's say one type of game. or maybe some very specific type of decentralized finance. I can see how that can do well, but they all end up needing the same kinds of services. And so they end up looking like each other and it's actually really difficult to compete against so many other layer twos, against other layer ones, and differentiating becomes very, very hard. And so to steel man the other side, it's really hard. Okay. makes me wonder, as the market kind of matures and realizes that layer twos are, it's not actually as easy to spin up an L2 and make it successful, build a community around it. I wonder if there's gonna be a revival in layer ones again, at some point. But be honest, if you think about it, technically speaking, there's a difference between layer two and layer one and so on. But from a community perspective, from an ecosystem perspective, from a distribution perspective, they're exactly the same. If you're building an L1 or building an L2, you will face the same specific challenges. Which is, how do I get my users? The building part is much more easy. It's quite easy. I know you can, maybe you can pay a million dollars, can get this chain built like, like if you pay a few million dollars, you'd probably even get chain link to get support your chain and you could, but then that would not solve your distribution problem, which is like, I'm building a DeFi thing. How do I get users to use it? How do I bring liquidity from Ethereum to this chain? So from that perspective, I agree, which is people realize that, you know, there's no real fundamental difference between an L1 and L2 from all those perspectives, which is- everything minus tech if you remove the tech yes there's tech differentiation but if you remove the tech part everything would look exactly the same which is you will have the same exact problems that an L1 will face or a new L1 will face so for example move megaeath you know some of the new L1s that are coming out or rollups that are coming out there's no difference between them and avalanche they have the same problem so in some sense I mean, okay, one benefit that these chains have is, or the newer ones have is, they could always argue that, look, know, yes, Ethereum is old, but it's not old enough. People, there is still not a killer application other than some of the financial stuff that we use, like Uniswap, Aave Compound, there's nothing else. Like people have tried front tech and it didn't work out very well. The only thing that seems to be working is pump.fun and they're... variations using AI agents, is, it's pure speculation, right? And so, I mean, one argument that they would make is, look, yes, I'm building an L1 and I have those challenges, but there's still a lot of application use cases that still under explore or haven't actually been found yet. And so whoever clicks or finds the first big consumer application will actually be the winner, which I agree. But there's nothing so far. Yes, there are pockets of here and there, things that you you could say, okay, here's another dex or here's another game, but there's nothing major that has popped up, which is also an opportunity. So whoever L1 and L2 can find that one thing, it doesn't actually need anything else, to be honest. Like imagine building an L1 and having let's say a twitter like thing running on that thing that's all that you need you would be the biggest someone by the way was saying that you know of course vcs don't fund daps anymore because they feel like there's not there's not much upside the values are low and so on but the guy who actually makes a really useful application and he actually finds traction he would be the star of the show for every single l1 every single roller would be billions from at least tens of hundreds of millions to this guy to come and build on the chain. And that is not there yet. So there's also an opportunity. So I would say, yes, infra is one component, but I think what's still missing is an application that... Like today, if you think about it, DocuSign or some of these applications, they're simple. There's nothing complicated about it. And yet you use it. Every business uses them. I think in crypto we have made the whole system so financial that, yeah, even thinking beyond an actual financial stuff and say, okay, here's something that people actually need. It's way too complicated and just not incentive to actually build some of that. And quite likely it's also possible that if you try to build that you actually won't get funded. And so, yeah, I'll come back to you saying that, know, yes, L1's... LL2s, they would look very much the same. I think the winner would be any L1 or L2, doesn't matter which, whether you're Rollup or LL1, that actually is able to attract a legit use case that is beyond your classical financial and speculative data. I agree. The experience that I am reminded of, or that reminds me of what I think supports what you're saying. When I led growth at Harmony, we saw a massive amount of assets, collateral that was being bridged from other chains to Harmony. at first we were wondering what's going on. And it turns out it was because of one application, DeFi Kingdoms. And it turned out to be a massive, massive magnet for users, for community. mean, it was immense. The growth that happened in a short amount of time was because of one application. And I haven't seen that type of application really since. I've seen a lot of kind of... I think to their credit, developers try, but, and I still don't understand the, I think the interesting about DeFi Kingdoms is it was really just kind of, it was novel. It was an interesting kind of game, kind of dynamics, a different skin to the DeFi that people were used to. And it drove on. a tremendous amount of collateral and assets and users. It was really amazing. And I can see how any L1 or L2 that has that type of application can bring that kind of users, I think would be a winner. I agree with you. Having worked with so many L2s in your time at Altlayer, what types of strategies have you seen that's worked to build community? That often becomes one of the hardest things to do, the harder thing to do. What have you seen that's worked? Let me give you, in our context, how we thought about this, right? So there two parts of this. One is like a retail community, a community that would support you for whatever you're building. It was very clear to us that, or it became very clear very early on that in this space today, the way it exists today, your community will not be there if they don't make money from whatever you're building. They have to make money. And you can see that, by the way, very recently with the hyperliquid, there's an immense support for this project. But if you actually look deeper into it, it's super unclear what the chain, by the way, this is an L1. It's a chain. And you have this text built on top of it. It's very unclear how this chain actually works. Some people say there are four validators and so on and whatever. But it's very unclear how the chain works. It seems like a black box. And there's a huge community. of people supporting it and now using it. And one of the reasons was because the price went up and people made some good money and they became an evangelist for that. We saw that with Link changing by the way back in the day. There were armies, marines and whatnot, people were calling them. It was all because of the token price went during a time when none of, I think it was in the bear market or something that Link was going up. So that's one thing, people don't realize it. I think the newer generation that's coming up or the newer generation protocols that are coming up, that are coming up, they're actually realizing that, you have to give, you have to make sure that your community actually makes money from whatever you're doing. That's one component. The second thing that people are realizing is that it has to be engaging. And you can see that through, I guess, the... engagement that MegaEath or some of the other projects are trying to achieve, which is engage your audience in a non-super technical way. Of course, they are technical founders and they will talk about technical stuff, but there's like a, you have to engage them in a slightly, or there has to be other team members that should engage with the community in a slightly non-technical manner. So what people call vibe in some sense, know, Bernard Chin is a good example of this, right? So there has to be something beyond just tech. That's one thing. And for the tech part, it's again very obvious that it has to be founder led. You can't build a community when your founder is not involved with the community on a day-to-day basis and is not fighting, debating, and trying to assert that whatever is building is the best thing out there. So I think we have seen that with many founders out there. think those founders have been quite successful where they go out there, defend their product, and do it regularly over time, at a point where their persona becomes established, and then the committee follows them. Because sometimes you, as a community member, you join a community when someone else who is smarter than you joins that community. So if someone else talks about it, you actually start reading more about that project as well. So I say there are three or four things. But again, you can go deeper into this. But I would say one is your company has to make money. And people who have screwed their community either by doing weird airdrops or putting weird. mechanisms of claims and so on, have, their community have, you can always see the pain that community and you would always see that the community is sidelined or not happy or not engaging and so on. And then the founder has to be engaging and then there has to be some fun element into it. People are bored with, I mean, not everyone understands all the tech components. So you need to have, find ways to engage your community without all that, whether it be your. mascot or whether it be a bunny or whatnot, people are all sorts of things. And sadly, shitposting has become... to be honest, you have to be able to entertain your crowd in some way. And someone was saying, if a competitor is doing it, what's the excuse that you wouldn't do it? Right? Everyone is doing it, so you should do it as well. But again, it's again the point that you have to be able to find your own voice, be able to engage your community in a way that stands out. And that's not obvious, but you have to find your own way, I would say. Now, going back to AltLayer, your primary customer are developers as well as app developers and also, I guess, people that want to create their own chain. What does community look like in your world and what's your approach to the various segments that you just mentioned in terms of community? the technical audience, there's the retail audience. What does that look like? So we have two, so we operate slightly different compared to an L1, L2. And here's the main difference. If you're building L1, L2, you have to find developers yourself and you have to convince them to come and build on your chain. You have to tell the world that, okay, your chain is the best chain out there. You have the best tech, you have the best community, best liquidity and whatever. And then they come and build on chain. For us, it's different because we don't have our own chain. We basically help people build their chains. And so initially we were, okay, by the way, we also had our own stack back in the day, a stack that before optimism was even a thing. And basically it allows you to spin up an optimism light chain. We abandoned it, I think, after six months of development because we felt like, after Optimism got released, we basically abandoned it because we felt like it's going to be very difficult for you to compete with Optimism when Optimism is already established brand and name. And so what we ended up doing was we started to build relationships with all these stack providers. So Optimism, Arbitrum. Foundation and ZKSync, Polygon, CDK, and Scrawl and whatnot. And our goal with that relationship was to be able to use their BD pipeline because we wanted to become their distribution partner. And so our idea was that, we don't have to push or go for or find developers ourselves. We know that they have the incentive to push. Optimism has the incentive to find developers for OPStack. Arbitrum has decided to find developers for OB stack. That's their business. And if that fails, of course we will fail. So there's like, they have to do it. They have to make it work. And as long as they need us, we'll always in business, right? Because they have to push their stack. And what we realized that over time, all these providers, they realize that they cannot handle the incoming inbound requests from developers because developers used to come to them and say, I have this question. How do I customize this? How do I customize that? it was too much for them to handle. And so they wanted someone, some implementation partner that could handle those requests for them. And this is where we came in. And so initially we felt like we had to push BD ourselves, but we literally stopped BD ourselves. We have literally a two-man BD team for a long time. And so most of the, I guess, that were coming to us or clients that were coming to us were through those contacts. or through the BDPy plan that the underlying stack provider had. So that is our kind of community in some sense, or whatever you call it. That is the base client base that keeps the wheel turning on our end. Now, we knew that that's not something that would directly involve our token holders in a way, right? Because yeah, we can spin up a chain for people, but that said, it doesn't directly impact our token holders. And this is kind of where we got involved in the Agile ecosystem, where we said, look, all these roll-ups that we're launching, they will need to find their own, they need to find a way to get their token involved. And if there was a way in which we could get our token involved with all these rollups, then we have a major use case for our token as well. So imagine we have 40 chains that we launch. And if each of these 40 chains, we have a way to insert our token in some way, then that would be magic. And so that was the goal with AVS that we started building, where the goal was we will launch these rollups, and for each of these rollups, we'll launch an AVS, and that AVS will have a token component from our side. And then anyone that has all token can stake their all token to be able to support and secure those AVSs. And in return, they will get rewards from us, but also from the roll up provider, so the roll up owner. So who owns that roll up. So this is kind of where the community got involved. And they felt like, okay, this is where it's not just a close and vendor locked RAS service, but it's a much more community oriented. involvement that they felt they connected to. Okay, this is really interesting. So there's two questions I want to follow up with. So it sounds like your BD pipeline is you become a distribution partner to these layer, layer twos, where developers want to build, let's say on scroll or ZKSync. They refer the developer to Omrit and AltLayer. Hey, these guys work with these guys. They'll get you set up really quickly. And so it becomes an inbound referral from the Layer 2s. And just to be clear, today, even if you want it, unless you're uni-chain, unless you are like the biggest out chain out there, then probably they will give you support. But if you are any other chain, they would just steer them towards one of the, not necessarily us, it could be other RAS providers, but they would steer you. would strongly say, we're not gonna answer your question. Go and talk to a RAS provider. That's the initial response, which is rather good for us. But this is how we tap into their BD pipeline. And that's good for everybody. It's good for you. It's good for the layer two. And also for the developer that wants to build in that layer two, because they want to get set up really quickly. you guys, AltLayer can do that. Going to the AltToken, we don't typically talk about kind of token mechanics, but it sounds like you've created a very interesting feedback loop with the token and aligned to the L2s that you have helped kind of set up. What is, can you maybe describe what the mechanism, what that looks like and how it works? Yeah, sure. basically, people stake all token and what they get is what we call a reall token or restake all token. Basically, it's a token that carries staking rewards from staking that all token. And then reall token can be used in the AVSs that we support. So you could use reall token to stake into, I don't know, let's say a Mac, a fast finite AVS that we run for dodo or so on. And then ReAlt is also supported on Eigen DA. So you can actually take ReAlt and put that into Eigen DA as a supported asset. And then you could get rewards from Eigen for supporting that Eigen DA. And so this way we basically support every AVS that we launch. And ReAlt takes a part in those AVSs. And so anyone. that is taking re-ault will get reward from also from us, but also from the underlying roll up that is being served by that ABS. And so as a layer two developer, I would have a greater incentive to work with alt layer versus competitors because I have an upside with the alt token in addition to the rewards from the underlying AVS that working, that is either that I either support or is being spun up on my chain, something like that. There are two benefits by the way and this is very unique to us. If you look at all the RAS providers out there, they are like, know, Caldera, Conduit, Altair, Gelato, I think these are the four major, Alchemy now has come up, the four or five major RAS providers. And there only two of them that has a token, Gelato and us. None of the others have a token, yet. Maybe they will launch at some point in the future, but as of now, They don't have a token. What we realized, by the way, is when we launched our token, the pipeline changed, by the way, because people used to come to us previously just for the tech. can you help me with the infra? But now they want more than just tech. They're looking for advisory. OK, now that you have a token, or you have launched a token in the past, can you assist us? Can you help us how to navigate through that process? So it has become a lot more. beyond tech and say, we will have a at some point. Of course, everyone will have a token at some point, but how do we play that? And Altlay at that point, I think even now today, Altlay is well positioned for people to advise you on how to go about doing that. So that's one angle. The other thing is by having none of the, by the way, other RAS providers have this AVS angle. So, and why, by the way, why didn't we have the AVS angle? Of course, know, Alt was a play, but also because when we talk, when we're talking to the clients, they always have this feedback where they said, look, I have this token and I will have a token as a roll up builder. Where do I get my token used? And the only thing at that time, that you could do with your token was put it as a gas token. So you take your token and you can use it as a gas token for your chain. And initially it felt like the best thing to do because then it kind of becomes like an L1, right? Because you're charging fees in your token. So it seems quite logical that it will be a, it's a good idea. But I think very quickly we realize and also clients realize that it's a really bad idea because people are used to Ethereum, like we are in the Ethereum ecosystem, so people are used to using ETH as an asset. So Optimism doesn't use any gas token, it's ETH as a gas token. Arbitrum has a similar scenario, BASE is similar. Now, you being an app chain, why would you need your own gas token? It didn't make sense from a user experience perspective, because now user has to and acquire that token to be able to transact on your chain. And so it was a painful user experience. Second is, if the idea that was that, OK, by having a gas token, I can actually capture value because people are transacting in that chain. That was also bad idea because in the end you are a roll up and so you have to post transactions on Ethereum and for that you have to pay in Ether. And so at some point if you're collecting fees in your token, at some point you have to do the conversion, have to do the exchange. And so you're basically selling your token for ETH and that creates a constant self-pressure on your token. And so it didn't seem very logical to actually do that. And the third thing, third reason was that if you have your own gas token, it means that your interoperability becomes more complex because let's say you have your chain with, let's say, Amrit token, and then you have a chain with Pete token. And then now, in order to do this interoperability, each side has to own each other's token, which is kind of weird. And so all of these reasons kind of convinced us and also the clients that using your token as a gas token is just a bad idea. And so what is the other alternative? And this is where AVS comes into play where you can actually build an AVS that would support your roll up as let's say for example, for decetalized sequencing, for a decetalized challenge in network and so on. This is where you can get your token used as a sync. So you can take your token and say, okay, if you want to become a part of my decetalized sequencing network, then you have to bring my tokens, take my token. And so that becomes actual sync for you to get, which is otherwise not possible, know, kind of classical model. So I think that's also a benefit that people saw with our kind of suite of products around AVS that have been built. Well, it sounds like the developers are really benefiting from a deeper partnership with AltLayer. So it's not just on the infrastructure side, but also on advisory. And I think that's super helpful to these developers and adds more value to them. And you guys get to do that because of the experience that you've had. Now, I want to get to an announcement that you guys made. And I want to read this announcement and get your feedback on it. I'm going to ask two questions about it. First of all, what does the announcements process look like at AltLayer? And I'd love to learn more about this new feature that you guys are providing. AltLayer has officially launched Autonom, a platform to facilitate the development, deployment, discovery, and distribution of verifiable AI agents that not only think, act, and make independent decisions, but all their actions are publicly verifiable. So first question, sounds super interesting. Would love to learn more about this new feature that you guys are providing through AltLayer. And also the announcements process. Now every project kind of has their own process of how they make announcements and distribute the message. What does that look like at AltLayer for you guys? Yeah, so then we have two, mainly two kinds of announcements. So announcement that is for our own products, so where we don't need anyone's external approval, so it's all our thing. And then the second kind of announcements are where we are announcing with a partner, for example, let's say, hey, we working with Optimism on launching this chain or whatever, right? So let me focus on the second part, which is our own internal announcements. So we, of course, we start with a blog post because it gives us, half of the time it's me writing it. So you basically start with a structure that tells you, what you actually want to announce. So usually it's a very long blog post. So it can range between five to 15 pages because what we end up doing is we use that announcement sometimes as our internal documentation in many cases. And then what we do is we start with that long, let's say five to 15 pages blog post, and we then stream it down to like a PR post. So post that would go to, let's say a journalist and say, look, here's a new product that we're building. Here are the main things that you need to know. And so we kind of distill down to like a smaller one page or two page or kind of announcement. And then we kind of put an embargo on this and say, okay, we want to announce this on this date, on this time, and here's an embargo on this. And then alongside we prepare our like tweet thread and know, discord messaging and so on. And what we also do is we, and by the this has become quite popular in the crypto community. This was, I didn't see this happening, I don't know, two years ago, but recently what you do is you announce this and then you also. make this you inform all your partners your vcs and you know everyone that you think should give it a like or a retweet or a co-tweet so you inform all your partners that hey this is coming out in the coming hours please make please show your love and then this by the way engages it also helps with tutorial algorithm so you know the more encouragement you get it lifts up as well. yeah, so we push that out and then it becomes public. And then we, of course, we through our own personal handles, we kind of amplified as well. So I would say, yeah, this is kind of the standard way of we do things. Awesome. And that sounds like a pretty standard process that I think most projects follow. Tell us about Autonom. Am I pronouncing that correctly? Autonom? a French word called for autonomy. So, that's a correct pronunciation, autonome. So now that this platform's out and it's part of the Outlier suite of services, tell us what kind of audience developer that would be interested in this and what have you seen so far in terms of interest and usage? to give context on why we started this, I'm not sure if you remember there was like a blog post or I think some work done by the Flashbot guys on a project called Teeheehee. Do you remember that? It like a month ago or so. no, I'm not familiar with it. Okay, so basically the idea behind that was that look, you have all these agents that are being built, like AIXPT and some of those, some of the really like Goat, for example, right, and so on. Like how do you actually know that it's an agent, it's like an AI thing. How do know that it's not a user basically typing all these tweets and pretending to be an agent? And this is by the way, very serious. I mean, imagine for example, one of these agents, by the way, I do feel like AIXBT, which is an agent from virtuals, it has kind of like a KOL status now where... I do feel that there will be an agent very soon, which will become so big that it will kind of create pump and dump kind of scenarios for tokens. So imagine a QL, like you buy this happens today, right? Imagine a big QL, the shield tokens, right? And say, you know, I have bought this token or whatever after buying the token, right? And then everyone kind of buys this because this QL has bought it. And... I won't be surprised if that rule will get replaced by an AI agent, or at least will have an AI agent version of that KOL. And the risk that I foresee is, imagine I am the developer of AIXPT, and then I want to pump my own bags in some way. And so what I do is I basically tweak the parameters and tweak the data that has been fed to AIXPT initially and say, you know what, you have to pump my bags and forget about... talking about Solana or whatever, just pump my thing. And this is where people who are believing or trusting that agent will get fooled because, you potentially will have massive financial losses if you have no proof that the agent is indeed trustworthy and it's not being biased by the developer himself. And so the idea behind Autonom was to build agents that would somehow be verifiable. So you as a user, would know which kind of agent you're tracking with and that no one has modified or tampered the underlying agent model. And by the way, there have been some attempts in the past of achieving this, but for some very simpler models. So there were things like ZKML, last DEF CON, there was things like OPML. which basically gives you a ZK proof. So ZKML, they gave you a proof that this agent was the one that was run for your inference request. But the problem with ZKML was that it only works with very simple models. It doesn't actually scale very well with some of the larger models that you use today with Lama and O and whatnot. And so it has those limits. Similarly, have with OPML, which is like optimistic version of ZKML, they had very similar limitations as well. It's not very scalable, it's not very trustworthy. And so what we felt like one way to do solve this problem was to use trusted execution environments, also known as TEE. And the idea is like, these are hardwares that you can trust and know that these hardwares are not manipulated. And so you then put your code in a secure enclave, which is a kind of a specific part of this machine that is protected from any outside interference. And then the hardware will give you proof that the code was not tampered with. And when it gave you an output that it came from this program that was stored in that specific enclave. And this way it will give you a proof that the agent has not been modified. And this is kind what we do with Autonom. So you take your agent, you deploy that, we take that agent, deploy that in a T, and then the T will give you proofs, depending on how frequently you want that, that the T, that the hardware is indeed running the same agent that you're expecting it to run. So it's much more scalable. You can run any agent you want. And so this is kind of what we're doing with agents. So it gives you very high ability on that, you know, the agent KOL that you're using is indeed the agent it has not been modified and tappered with. Is the the training data also in a secure enclave? Because that's that's the I think the part that is difficult to I guess to verify. we are not hosting the model itself. So there are two components. There's something like the fundamental foundation models, like. GPT and so on. This is where training happens and the model gets created and all the weights are handled and so on. And then there is agent that is built using that model. So for example, today you have this GPT model or the GPT different versions built by OpenAI and they expose that to you through an API. And you have an access to the API and say, I'm gonna use this API and then will access, make a query to that model and then the model will give you feedback. So you're not handling the model cells. So we have one abstraction, one layer above model. So you have model below and then you have the agent like as an application on top of the model. So we are hosting the agent, not the model. At some point in the future, we might do it. But so far we have seen is, of course there's some agents, some models that are open source, you can actually use them and run them in a T, but most of the models that actually being used today in the crypto space are agents that are not open sourced. the only thing that you can do is you can prove that the open AI API key was used. That's it. You can't prove anything more than that. But yeah, at some point in the future, we might actually host the model self, which is something that we can consider in the future. Yeah, as I think about these models, I think model enhancement is really how these agents can differentiate from each other. Otherwise, they're all working from the same training set and they all look the same and they all answer kind of the same things. Yeah. give them a bit of character, things like that. But it doesn't change the fundamental model. That still remains the same. You're just feeding them a little bit of feedback. Yeah, to your point about bias and you know, if I were the agent creator and I want to give them, I want them to bias answers towards the token that I already own so that these agents as KOLs will shill my token. I can see that's where the potential danger is, is in the model enhancement and pushing my specific data towards the token that I own. And so. thing you would imagine, imagine let's say you build an agent that helps you with, I don't know, financial stuff, let's say, don't know, trading or whatnot, right? And for some reason, let's say the agent, so I distribute this agent and somehow I inserted something that looks very fishy or that's very fishy in some sense and that you want to be able to check that. And that's interesting because imagine a DeFi app, right? The DeFi app is hosting or using an agent underneath for end users. And any users want to make sure that when I'm closing my position or opening my position, the agent is actually suggesting the right thing, not the stupid thing. And how would that happen? Only if you can actually verify that the agent is actually the same agent that you initially wanted to interact with. And so that is what DE gives you, this proof that whatever agent was initially deployed is the same agent that is being. invoked when you were actually asking them asking the agent for inference or request or what have you. Interesting. Well, Amrit, thank you so much. Any final words you'd like to share? Tell the audience of developers about AltLayer and how they might be able to reach you. of course. we have expanded our services from RAS to AVSS to now we're exploring, as Peter mentioned, we're the agent space as well. If you're a and you're thinking of, if any of this kind of makes you feel, it's too much for me, then come and talk to us. This is what we do. We will help you go to the market quickly. and you would be able to rely on us for all your info services and all you have to focus on. Okay, I have this info handled, what do I do next? I build my app or I go to market. So come and talk to us. We have, and I have my DM almost always open, so ping me on Twitter or ping us on all clear official handles and we'll definitely would love to help you out. Amazing. Amrit Kumar from AltLayer. Thank you. Thank you, Peter.