Vibe Code a Sports Score Oracle AVS on EigenLayer by Nader Dabit

Transcript

So my talk is really going to be about three things I guess maybe four things. Uh number one I'm going to do an introduction layer. Number two, I'm going to do an introduction to a tool called Waves, which is an abstraction on top of IGEN layer that allows you to build services verifiably a lot easier. And um number three, we're going to be looking at kind of like a vibe coding prompt that I created to do a oneshot creation of a service.

And number four, we're then going to try to live life code that uh using clawed code. And I would give this about a 75% now, let me say, let me be more. I would say I would give this about a 50% chance of working. Um because I'm live coding and demoing and in front of people. Um I've tested this prompt a bunch of times and it's worked about 95% of the time. So I'm giving myself some leeway, but uh hopefully uh hopefully it works out, you know. Um, and then one thing I want to actually do to modify the uh the presentation is by just uh bringing up this this visual here which I'm going to touch on

in just a moment. Okay, cool. So to start the presentation off with Ian layer, I'm pretty sure most people here kind of know about layer but a lot of people um maybe don't have a very great understanding about it. So part of what we're doing right now is uh trying to really really really dive in and polish and dial in what we uh use or the language we use to kind of explain layer. And part of what I want to do here today is kind of along those lines. So when people think about layer a lot of times they think about restaking because when I layer was first created there needed to be economic security that would then be the means to the end

to build what the original vision was. So I guess an analogy of what you might kind of think of in a sense of like you need one you need one thing to exist before the other thing to exist. But with IGEN layer, we needed to have pulled security in order to make it available for applications to build. It would be if someone like forked Ethereum and they launched, you know, another version of Ethereum, literally just forked the codebase and they have a couple of uh validators running or something like that. And then they like they went to a conference and they were like, "Hey, we you should build your application on our blockchain, but there's no security for that for that application. So why would anyone do

that, right? Instead, you would probably need to first validate your idea, bring on uh validators, literal literal validators, launch a token, have investment, have money, and then that valuation of that token would then provide the security for that network, right? But you wouldn't start at the end, you would start at the beginning. And that's kind of like how restaking became uh known as as part of IG layer. But ultimately what we're here for is for building developer platforms and essentially unleashing innovation and making it easier for developers to build applications.

So when you think about developer platforms, there have been quite a few that have really really gone from like either zero to one or a 10x improvement. So when you talk about software or any products and you talk to an uh a VC or someone that is going to be investing in your company, they often tell you that the thing that you're building has to be 10x better than anything else that exists before for you to be able to kind of pull people away from those those previous existing things or they need to be a zero to one product. Those are kind of the two main things that venture capitalists are looking for to invest in your your product because if it's just slightly better, it doesn't really make uh a difference to the point where you're going to be able to kind of like

have people lo leave another platform to come into your platform or leave a product because you have that lock in, right? So if someone wanted to build a competitor to notion, it would have to not only have all of the features that notion have, but it would need to be like 10x better because a lot of us are using notion or we're using air table. We have these products that were like really ingrained into our workflow and if it's just slightly better, it makes no sense for us to leave, right? So what are some examples of developer platforms that really move the bar? I think uh the first one would be the iOS combination of the app store Xcode and the um I would say programming languages Objective C and Swift. If you wanted to build a mobile app before this existed,

it was almost impossible. There were probably more people like I I would say the same amount of people maybe in this conference that knew actually how to build mobile apps um before the app store existed on this entire earth. like maybe like aund 150 maybe a few hundred people that could build a mobile app before these uh actual platforms existed. What Apple did was they realized okay we have this device everyone has it in their pocket. How can we make it to where we can unleash innovation for anyone to kind of come up with an idea and make it available for them natively on their device. So they decided to build not only the distribution mechanism which was the app store but

they also had to actually build the developer tooling which is Xcode and then the supporting programming languages that they either built or supported Objective C and Swift with the app store and in Xcode you could download Xcode you could start with the starter project you could literally write a few lines of code and launch an app to the app store and have distribution to all these people. This just really changed the game and it made a lot of people excited about software development including myself. In fact, the reason I decided to learn how to code was because there were so many people building and launching apps. I was like, "Oh, I want to build an app and it's so cool like I want to learn how to code." So, this unleashed innovation for mobile mobile applications. You also have a cloud. So before the cloud existed, if you wanted

to build and scale an actual uh service globally to tens or hundreds of millions of users, again, there were literally probably a few hundred or maybe a few thousand people on the earth that could actually show you how to do that. um because you needed to actually do a an an insane amount of work to make your service not only stable and reliable but also um I would say with enough like low latency around the globe to make it usable and uh and therefore it was just really really challenging. With AWS, with Google Cloud and with Azure, you can actually now spin up databases like DynamoB that handle tens or even hundreds of millions of operations per second. You can spin up uh services like uh on AWS Lambda that will autoscale for

you and scale down depending on how many users. So it makes it trivial. The only thing you have to focus on now is writing your for for the most part like your business logic like what is the actual differentiating product you're building. You don't have to worry about what they called undifferiated heavy lifting. So this paradigm is how 95% 99% maybe of software companies build now. So this is an an insane amount of innovation that's kind of been unleashed. And then you even have developer platforms on top of developer platforms. You have Verscell, you have uh you know cloud cloud flare um and and these other abstractions.

Verscell is actually a customer of AWS. So when you're using Verscell's functions, you're actually using AWS Lambda under the hood to the best of my knowledge, at least the last time I checked. I know they've released a couple of new products lately. I don't know their their infrastructure. But with Verscell, I can go and I can download a Nex.js project and I can like type in one command in my terminal and it gets deployed to this global CDN. Everyone in the world can use it. Automatic CI and CD. I upload my code to GitHub and make a change. It gets redeployed. It's just an insane like 100x better uh improvement than what we maybe would have done 10 years ago. We might have written some code, compiled it, upload it to like an FTP server,

right? And we we wouldn't even have had like a CDN or anything at that point. So anyway, so these developer platforms, they really unleash innovation. So I can say, I'm going to go v code an app now. I'm combining AI with Versell. I can upload it. I can get it in front of uh the entire world. Low latency, it scales. Boom. Everything is ready. So in the blockchain world, what is the analogy there? Well, it's basically smart contract blockchains or block programmable blockchains. That's kind of where we are at the moment. Before Ethereum existed, if you wanted to build a blockchain application that was different, you literally either had to fork or build a new blockchain from the ground up.

Bitcoin, Litecoin, Namecoin, there were dozens, maybe hundreds of these uh slightly different networks that existed before Ethereum, but you couldn't build an application on them. You had to fork them and then change something and then you had to bootshop that whole process from the ground up and obviously there was very limited innovation there because it was just a hard thing to do and very limited number of people that knew how to do that. With programmable blockchains you write a smart contract. You can literally take an existing contract, make one line of code change and you now have deployed a brand new application and it takes 5 seconds. But what you're doing with Ethereum or programmable blockchains is that you have all this underlying pool of economic security that you're

inheriting. You don't have to go and start a new network. You don't have to launch the ETH token. You don't have to do all this stuff. You're kind of just launching an application. You're inheriting that economic security. Right? So with that in mind, that's a good analogy of kind of how how I layer works. And I'm going to dive into um yeah, let's dive into a little bit more there. So with IGN layer, turn this slideshow back on. There we go. So there we go. So we have like the application specific blockchain paradigm. You build the network from scratch. You have now programmable blockchains where you can write a smart contract. You're inheriting that underlying e economic security. And then

we had we have the rollup era now. So you can kind of like modify your execution, but you're still under you're you're still inheriting the underlying consensus and security. So you can actually deploy an SVM or a move VM or a different EVM to this underlying network, but still inherit all that security. But this is extremely limiting even at the programmable rollup stage. Why? Because we're building within the constraints of a blockchain virtual machine. If you're building on Salana on SVM, if you're building on move on the move VM on Optos or Sui, if you're building on Ethereum or or any EVM, you are constrained in the sense that you cannot make external API calls. You

cannot do complex computation. You cannot do GP like complex GPU. Um there's just you can't even talk to other blockchains, right? You're just like extremely limited. And people ask like why the hell is why is innovation been so stifled in the crypto space? Why haven't there been killer apps? Well, we're we don't have much to work with. We're just extremely limited on what we can do. So the idea for layer in general is essentially okay we have all this underlying economic security but we're constrained within this network. What what are people building already that are kind of like already expanding beyond a smart contract? Well, let's talk about I guess uh oracles are a good example. If you

want to talk to the outside world, you can do that from a crypto uh like I would say from within a uh crypto v uh virtual machine, right? If you want to for instance say get a price feed from um from some oracle like uh you know chain link or something like that. The way that you do that is you you have an event that's emitted it's picked up offchain something happens and then it sends it back on chain. Now how was chain link built? Well chain link was built like Bitcoin was built. It's literally a completely new network from scratch. So if I'm like okay I like chain link but it kind of doesn't do everything I wanted. I want to build something slightly different. The task that you now are presented with is

literally building that entire network from the ground up, creating a token, bootstrapping validator set. It's an insane amount of work for one little piece of functionality, right? So what if we could instead kind of take the pulled economic security that we have from blockchains like Ethereum and Salana and then we can combine that with any arbitrary off-chain code that we want to write in any programming language TypeScript Rust Go Python whatever and then we can kind of combine the two that's essentially what I layer is or really what shared security ultimately enables.

So if you want to say build an oracle, if you want to build some type of verifiable inference from AI inference, if you want to build some type of uh I don't know bridge, if you want to build a co-processor that does some insane complex computation that isn't possible within the virtual machine, if you want to talk to an external web uh 2 API, I mean literally anything you could possibly want to do is now instant and and very simple in theory. In practice, I'm going to talk about how we're kind of making it a little bit easier. But now with IG layer, you can say, okay, I'm going to emit an event. I have this JavaScript. I'm going to pick up the event. I'm going to do all this stuff and I'm going to send it back on chain. But but this offchain code no

longer needs to then be secured by its own separate validator set, its own token, and all that. I can just go do whatever I want. And then I'm making this uh possible by inheriting this other uh underlying economic security. Now does that mean you still don't launch a token for your for your service or your software or your app or whatever? It doesn't it just depends um what would be the uh different like circumstances there would be okay if you did launch your own network from scratch and you did launch a token you're probably paying tens of millions of dollars or even like depending on how like valuable your network is millions or tens of millions of dollars per year

in emissions for the validators that are running your network. you you can pay a small fraction of that to the operators running your ABS. Like when I say a small fraction of that, we're talking about, you know, much much smaller fraction in the range of like singledigit percentages of what you would have paid in emissions. And to kind of visualize this, I created this uh diagram a couple days ago that I thought was like pretty pretty good because it kind of really really uh puts things into buckets that kind of make sense to a lot of people. So you have application specific chains that are very powerful but they can only do one thing. you now uh then to expand the scope of that you have smart contract

blockchains or programmable blockchains that allow you to write your own logic and do everything but you're restricted by the VM that you're working inside of so what can you do here token transfers you can store and track ownership of assets you can do uh you can do I would say trivial database and data computation you can verify signatures and you can authenticate users you can interact with other uh uh contracts on the same network and you can basically build uh within the constraints of that virtual machine. With IGEN layer, you can literally take all of these things that you can do within a blockchain because you're still going to be working within a network from uh you know to to start with and then you can expand that with anything that you can potentially

think of that you would write into a a traditional web two server. So I'm going to skip over well I guess I'm not I shouldn't skip over let me just show this this slide here. Um this is just an idea of like all the different players within the IEN layer uh marketplace. So I layer is kind of like a four-sided marketplace you can think of. We have the restaking which we talked about at first which is kind of like where the economic security comes from. You then have the ABS's which are services. AS is an acronym for autonomous verifiable services. You can think of these as networks. In the past, people used to say chains, but these are not just chains. They're like all types

of other things. So, you have the security, which is the restaking. You have the ads, which are the services. You have the stakers, which are the people actually like staking, you know, the the rewards. You have um the operators who are actually running the ABS software. So, stakers would be the economic security. Uh ABS's are the services. Operators are running the services. That's the third side and then the fourth side would be the consumers. These are the applications that are kind of consuming you know these these services. So we started with restaking we then moved to ABS's and we've been focusing on ABS's which are services for the last year. We now have uh I think around 6 to 700 that we've talked to that are like either exploring or

building and we have around 200 that are building towards mainet confirmed and we have around 50 or so that are already on mainet. Um, so with that, ABS's we're that's continuing to gain gain momentum. We were the fastest growing blockchain developer ecosystem in the world in 2024 and we're hoping to continue that trend in 2025. We're now starting to see more and more applications being built on top of these services. So 50 live services, 200 hopefully coming within the next 6 months. And now we start to have like actual rail applications. A couple of examples of those would be um Earnify which is I think it was the number one app in the app store for finance. We have um NOS um we have the new the founders of uh Reddit and DIG are

building the new DIG uh on on opacity network. Um Redstone is like an oracle that's built on layer that has a bunch of applications. I think uh when we kind of last tallied it, we had around 80 what I would consider like notable applications, but by the end of 2025, we're hoping that number to be closer to like a thousand. So that's a huge focus for us is to kind of when I say like notable apps, I'm not talking about like a hackathon project. I'm talking about like an app in the app store. And then as more ads and and more apps are live then the rewards are being paid to the operators giving them higher yield which will then attract more uh TDL right so you kind of have this equilibrium where um if the people that

are staking or earning rewards then they stick around if they're not they leave but when they leave the yield goes up right so the more ads the more rewards that are coming online the higher the yield the more the TVL goes up but even if the TVL goes down the people that are still there make more money So, it kind of balances itself out or more yield. Um, couple numbers, you know, staking went way up in 2024 obviously because it's our the year that we launched. Um, operators, we have 75 of the world's top 100 operators running services on layer. We're shooting for all 100 ABS developers. Yeah. Um, you know, we grew extremely well in 2024. We're looking to keep that going this year.

And what are people building? Um, different verticals are listed here, but a lot of this stuff is a little nuanced, like it overlaps. So, I wouldn't kind of say that something that's outside of the rollup services category doesn't actually overlap with that and vice versa. A big place that we're seeing a lot of people building products that are like fundraising and have some cool opportunities are in the AI space. Sentient Ritual, Hyperbolic, Gaia, um, Open Leisure, there's a bunch of cool ones that are being built in that space as well. And then we have IDA which is the most scalable data availability solution. We're about to launch our V2 or Blazer update which will um increase throughput along with a few other

things. Yeah. A few of the different protocols that that I would note that are building or already live on layer would be like layer zero um and poly market are building a prediction market oracle and fura is building um you know their RPC infra a new RPC infrastructure mega is building on huge fans of mega um mantle as well is one of the more popular oracles in the in the blockchain world hyperland is a cool bridge I think I mentioned some of these AI projects as Well, Aether is really uh really interesting because they actually have revenue. You know, there's not that many companies that have revenue in crypto, right? Uh this this number is wrong.

It's like $130 million annual revenue. Um Zerther has. All right. So, road map. Um I think we're actually already shipped some of the stuff that's listed here like slashing. So, slashing was in progress. It's now shipped. uh we're working on data availability throughput of 100 megabytes per second which would be around 75 times more scalable than the next leading uh option in the space and uh really lowering the barrier to entry for developers is a big focus for us. So we want to make building ads 100 times easier. Uh a good analogy might be serverless functions. If you write a line of code in Typescript and you want to execute it in the cloud, you can you

can deploy it to Lambda and it just runs and everything else is managed for you. If you want to build a completely brand new service from scratch, from the ground up, you can go the more complicated route, use Kubernetes, you can use EC2, you can use all these like more low-level infrastructure tools, but AWS gives you the option to do whatever you like. We are going to be offering the same options. We want to make it to where if you want to just write some code, deploy it to IGEN layer and let us run it, we want to make that make that possible. There are teams like Authentic, there are teams like Waves that are building great developer tooling that allow you to do this today.

Um, so a lot of like ways to kind of build on layer and um, if you ever want to learn more, you can chat with me, but I'm going to be diving into Waves today. So, all right. With that being said, um, I'm going to be diving into the de coding coding demo. Now, I think I mentioned uh I'm going to mention more a couple more apps. I think I mentioned Dig and Ernify and NOS. is interesting. Daisy is a web two social application using ZKTLS. Nsh delivery is really cool because it's essentially just a true web two application um similar to Uber Eatats and the number

one focus for me and for the company uh this year after obviously security is number one. Number two I guess would be developers. Um we are like really focused on uh bringing in more developers building apps and building services this year. Four ways to build an IGEN layer I would say. One build an an app on an IE layer AS. So like building on an existing service like opacity network um or reclaim protocol or whatever there's like 50 of them like I said. Number two would be building ABS from scratch. Uh number three would be using a developer abstraction like authentic or waves to build an ABS. So you could think of authentic or waves similar to versel on top of of AWS. It just makes it a lot simpler to to build

and get something to market. And they also have a great team that kind of like helps you out as well. And then finally, IDA is a separate product that you can build on that doesn't like uh have the exact same use cases as an ads. Obviously, it's just more of a data availability layer, but a lot of people are using it as like a temporary cache. A few different ways that people are using IGA outside of like what you might consider of a data availability layer would be used for. Um, if you want to get involved, we have the IGEN alpha ambassador program. Um, where we have hacker houses that are going to be, you know, continuing to pop up. We're going to be announcing one for Berlin very soon. Um we have a university program called IEN games and then we have a build on group chat that's open for

anyone that wants to join. Okay, cool. So now to the live coding demo. So I mentioned that the whole like ultimate goal is to make building offchain or bringing information offchain or being able to kind of like write offchain code and make it verifiable as easy as possible. We want to take the best properties of a blockchain which would be security and verifiability and we want to pair that with the best properties of building just traditional software expressivity. I can talk to a database. I can talk to an API. I can call an AI inference service. I can call I can uh work with some very very computational heavy GPU. I can do literally anything I want in the whole world. We want to combine those two

things together. So the tool that I'm going to be using today to make that happen is called Waves. Waves is essentially an abstraction on top of layer that allows you to just write your business logic. They have their own operators and everything. You just upload your code to their service and they run it on layer for you. It's kind of like the serverless functions of of IGEN layer. You could kind of think of it in that way. So the way that we're going to in and the the the programming languages that they support at the moment I believe are Rust and Go. I think Rust is probably the most supported uh programming language though. So when you start off uh a project with them, they give you kind of a few boiler plates. They have like maybe five or 10 really cool boiler

plates that you can start off with. But the two most I would say common uh ways that people start getting building with IGEN layer and two of the more common use cases are oracles and they are bridges. Um so an oracle again would be like I'm making I'm emitting an event for my smart contract that that says something like give me the final score of the Bulls game from today. The oracle picks that up and it goes and finds the score of the Bulls game because that's not available on chain, right? Then it pulls it back onchain by writing uh it's about like writing a transaction back to the network and then I pick that information up and now my cryp my blockchain application has that

information. So that's that's the general idea with an oracle. You make a call, you made an event, you pick that up offchain, you do some stuff and you bring it back on chain. But that could be anything like I can talk to any API, right? So in our case, we want to transform this oracle into an AI inference service. But you can't just do AI inference offchain because the whole idea of of having uh a blockchain verifiability and for verifiability, you need some type of um I would say determinism. You can do that in two ways. You can do that with uh ZK because you can kind of like verify that the outcome the output is truly what it is but it's a little more complicated.

It's more expensive and it's a little slower. Or you can use something like the approach that we're taking that basically uses uh a combination of a traditional like LLM service like OpenAI along with the seed. Now the determinism of this approach isn't 1 million% accurate. I think it's like 99.99. There's like some number of nines percentage that that this should work. Meaning that if you have like six or 10 or five, whatever number of operators and they all run the same query with that C, they should all return the same exact answer. So if I ask what is 5 + 5, every LLM will probably give me 10, right? But what if I say what is the best place to eat uh good uh creme

brulee in the world? I might get like 10 different answers, but but if you want determinism, you can actually make that happen by providing a seed for LLMs that support that, right? So, that's going to be the way that we approach it. We're going to provide a seed and we're going to kind of like take the Oracle service or we're going to kind of transform that into a verifiable inference. And the way that we're going to do that is we're not going to actually um we're not going to actually code this. We're going to vibe code it, meaning we're just going to send in this prompt that will just explain everything we want in one shot and we're going to let clog code write all of our code. And once you kind of like start working in this way, it

opens up the door for like a lot of cool right? Because you can then come up with any idea possible and then you can take this prompt and you can kind of modify it to build anything in the whole like world that you could possibly think of. And that's why I said like this this demo may or may not work because we're essentially um asking the the LLM to provide us with like 15 or 20 code updates for these files in one prompt and we're we're just like, okay, we're going to do all this and we're going to run the code and we hope that it works on the first try, right? That sounds kind of crazy, but that's that's that's what we're going to try to do. Okay. So to so to do that, we're going to create a new project and we're just going to

start writing some code. And um if you look here, I don't know if you can see that. It looks like you can. There's basically a a forge. You can use Forge to kind of clone a uh GitHub repo, right? So we're just going to use Forge and we're going to clone that that base Oracle project. And we're going to use See. So I'm just going to paste this here. So for net and we're just going to pass in the template being the waves boundary template which is an oracle. All right.

So now we have this uh this waves project. Oops. Actually I shouldn't have done that. Um all right. All right. So, we're going to open up this in our code editor. And this is Windinsurf. If you've never used Windsurf, this is what it looks like. Um, you have your typical like VS Code stuff on the left. On the right, you basically have your LLM like conversational prompt available at all times. Um, I'm not actually going to use Windsurf though. I'm going to use Claude code. So, I'm just going to close this.

I'm not going to use Windsurf's coding thingy. All right, cool. So, we have that. So, what we want to do is just build out the Oracle. And here here is one of the representatives or founders of Waves, by the way, which hopefully I I do it justice. Anyway, if you want to learn more about Waves, talk to him. Anyway, so anyway, I'm here. I'm going to go ahead and um try to run this basic service from uh from the template that was given to me. So, we can kind of just run a basic example and then we're going to modify that. And what this example does, it's kind of a uh I think it's a price feed for getting crypto prices. So let's say you want to know the price of Bitcoin and you want to bring it on chain into your application or you want

some random arbitrary coin that literally launched 5 seconds ago. You can do that. You can make that. It doesn't exist, but you can make it, right? Cuz obviously you're not going to see this supported by um chain link or or these other ones, but you know, if you build it yourself, you can build anything you want. So, to kind of get everything set up, I'm going to run make setup, which is just a make um command that you can actually go down here into the make file and check out. And what I would say is actually if you're building with Waves, the the first place I would go would be the make file because all of the different commands that that you can do within the project are all located in the make file. Um, and it also has a lot of like helpful information to help you

understand kind of the underlying infrastructure of waves. And when you look at this waves program that it's given us, it's actually very very verbose in terms of like how much boiler plate is there. But this is the case with a lot of great, you know, uh, software starter templates, right? If you've ever built with React Native, you you start a new React Native project and you have an insane amount of like dependencies, but you go into this one file and you write your code and everything runs beautifully. And that's the one place you need to worry about. You don't need to worry about all this other stuff, right? It's very similar here. We're going to go into the components and we're going to go into Oh, you know what? I'm going to actually delete this project because I think this is um a

project that I've already built before and and it's not going to actually Let me just I need to basically start over again. Just give me one second. Oh, you you you initialize a project that already existed. Yeah, that's kind of what happened. It's okay. It takes two seconds. Okay, cool. So, we're back into we're back into our project here. Okay, I'm going to go ahead and run um make setup again.

All right. And now if I go into the project here and I go to components, you'll see that we have uh this ETH price oracle here. Okay. So this is the this is the file. This is the little area that we're working with. And in reality, the only code you need to really start with is this lib file. And within this lip file, there are like two functions that you might even then really want to worry about would be the run function and then the function that actually calls the oracle, which is this get price view. I'm sorry, which calls the API that that that is essentially the main functionality of your oracle.

So this code is very simple to to kind of read and understand for the most part, right? You have this get price feed function that calls in this case the coin market cap API. It passes in an ID that's then in this case set in the make file and then it just returns that value in like a JSON object and that's about it, right? So not not too much going on there. So let's run this. And before we run that, let's go back to our make file and kind of like look and see where this config is set. Right here we have a coin market cap ID that is number one on on coin market cap API. Number one is Bitcoin and number two is Litecoin. So,

what we're going to do is we're going to start with Bitcoin. We'll get the Bitcoin price, then we're going to go and we're going to get the Litecoin price just to kind of test it out. Now, I'm going to run forge build and depending on the day, this could take a long time or not. Oh, wow. That was fast as hell. Okay, good. Awesome. Okay, cool. So, the project has been built and um Okay, let's let's let's actually not test it. What we're going to do now is we're gonna um we're actually gonna let me let me run this make wazi build. I don't even know if this is necessary, but let's just run it real quick. I think this might be the longer command.

Okay, so I haven't gone into kind of like how this is how this is actually working under the hood. So I I'll I'll touch on that in a second and I might not be giving it justice. Um so maybe you know again if you have any questions um Sam and the founders of like waves are here or layer waves whatever. So when you build on IG layer today, you're building um essentially a containerized often uh version of your application and then you have to find operators to then run that service and it's kind of a manual process, right? You find you have to coordinate with these operators, they have to download your software, they have to run it and then if you have an update, you have to then coordinate the updates. So it's kind of like, you know, a somewhat intensive task. With uh with

with Waves, they actually have their own operator set. So all you have to do is is worry about focusing on your core product or business logic. You run a command, it kind of bundles it up and then you can deploy it to their operators and then when you want to make a change, you kind of deploy that update again. It's kind of like again a serverless function. You don't have to deal with a lot of like of this uh lower level infrastructure. You're just like writing your business logic and deploying it. Okay, cool. So we've run make wazi build which builds that was component because that's kind of what it is. It's under the hood.

this web assembly. And now we can test it out. Actually, maybe we maybe that was the command actually. Let's see here. Okay, maybe the build was just the builds. Let's try this one more time. Maybe this is the command.

No, that's the command. Yeah, I think there you execute. Oh, there it is. Oh, okay. I see. I was confused because they're passing in the coin market cap ID as a as an argument. Okay, cool. So let me just run make wazi execute here and this will execute and you can it looks like uh you pass in the um you can pass in the um ID here or you can take it from the um underlying environment variable. Oh, I need to start docker.

All right. Now, Docker's running. I'm going to run this again. on the manual before process hard. What's that? On the dock is like a process like I'm running the same Are you running into the same issue? Yeah. And I scroll up. Oh, okay. Scroll up.

Yeah, this commands before I have Docker installed and running. I thought Let me just check my I'm in the wrong file now.

Ask your support there for It could be some I don't know what is this actually. This hasn't been put into me yet.

I I wonder if there's some problem with um Yeah. Uh, I feel like Docker isn't being recognized in my um environment variable or something like I mean in my in my environment. Let me just see if I can double check where I'm going to use spot code real quick and I'm going to try to um bob code to fix this issue and then um if not we might have to like give up on the on the live coding demo.

That was good. It's telling me to do the thing that

I've already So, let me try that one more time. And if not, I might I might give up on the lab domain. But I have a feeling it's I mean, I know for sure it has something to do with my own Docker issue instance. Do advanced advanc What's that? I think I found an issue. It's under advance. It's like a log for Oh sorry.

Um I think it's under docker advanced allow the default Docker socket to be used this. Yeah, I mean that's that's what I'm not sure though because do you think it would I've been using Docker for the last ever since I got my computer so I haven't had any issues. I'm wondering if you're saying the configuration for um finding it within my path you think it's like a path issue or I I think so. And uh so right now you're running docker right?

Yes. And there is under the settings like advanced. Oh, okay. You're saying in the actual Docker settings. Yeah, the UI. For some reason, I don't even see them right now. So, let me try that more. So, let me just uh see if I can close. I don't even see it running. No. Yeah.

Wow. Yeah, I think I might have to give up. Well, let me just wonder if that had something to do with this update I made recently with Xcode. I doubt it, but who knows? Bigger volume fire.

I don't think this is going to work anyway. So, I'm going to have to unfortunately just uh quit the demo because Docker not running is actually a problem because it's kind of a big obviously the big dependency for this. So, what I would suggest though is like if you just have Docker installed on your machine, it should all work fine. Um, if you just like look for uh David 3 GitHub gists, um, you'll see that I have all of my gists uh, lined up here for a couple of different prompts. So, I have a verifiable AI prompt. I have a uh, sports score oracle prompt. Um, I have a

few different prompts that allow you to kind of like create different types of uh of services on waves. So, starting off, you get that uh that coin market cap or that price seed oracle. I have a I have a prompt that can let you create a sports oracle. So, if you want to get sports scores and kind of bring them on chain for building a prediction market or something like that, I have an example for that. And then the other one I have is like the verifiable AI inference. So, um I guess if you do want to see a working live demo of this, I recorded a video of this and it's on my Twitter as well and on my YouTube, so you can check that out if uh it might be even better than this because then you can go back and like rewatch it and uh

and all that. So, um so I guess with that being said, that kind of wraps up my talk. So, um thanks everyone for checking this out. I apologize that my uh my live coding demo kind of crashed out which is you know um unfortunate but it's pretty common honestly it's like it was Docker's fault. Docker crashed out not you but um but anyway so if you have any questions yeah come talk to me or talk to Sam who's also in the building and yeah he's waving his hand there. He's one of the founders of Waves and um I hope you enjoy the rest of your week here in Dubai. Thanks.