AVS development cycle by Yarden

Transcript

Um, okay. Thank you guys. Thank you. Thank you for having us. Thank you for being here. I'm Yan for CEO of Authentic. And today I want to talk with you about this new par shift, this new par of services. But first, I want to start off with a quick mental model. In the recent deals, we have witnessed the rise of different pieces um from application specific to general purpose chain from monolytic versus uh modular and of course these handling approach.

All those comes with different tradeoff but they all serve the same use cases which is building blockchain frameworks like SDK like blockchain mobile builders with the ability to build application specific blockchain essentially a framework of modular components that builders can configure and on the other hand L2 frameworks like Obstac They provide builders with the ability to build rollups which again essentially are blockchain even with improved transaction cability, higher LTPS, lower gas costs but again blockchains. So L1, L2 frameworks they

provide the ability to build blockchain. Those blockchains they come on to consensus of the new state of the network, the new state of the ledger, right? All transaction execution. Let's come to consensus. Let's come to state to state the agreement to state transition. And this is extremely limiting the potential use cases that can be built in the web three space because those blockchain they cannot agree they cannot come to consensus on offchain execution of offchain services of arbitrary execution again extremely limiting the potential of real world use cases that

can be built in the web3 ecosystem. So what is the missing piece? Okay, what if we want to build our boost offchain software? What if you want to build high performant software, resource intensive software on decentralized foundation at scale? Okay, what if you want to build organs? What if you want to build data labeling service for AI engines? So in memory fast databases, GPU clusters processor or even distributed network to train AI model, you won't be able to do so with L1 and L2

frameworks. So we want to build a service that run offchain that has unique logic, unique execution that has intensive computational logic. Okay, we are in and the idea is to build real world use cases. The idea is to build great apps but great apps they require great offchain middleware services. Let's take some physical examples. Okay, bridges. All bridges has offchain components. Okay, companies like Walmart, company like companies like zero, they all have offchain components but they are not blockchains

allocing have no underlining blockchain system as part of the wider infrastructure, right? Chaining is a network of nodes that come to consensus on some offchain execution and offchain computes. Okay, indexing services like the graph, they have no blockchain system within the wider infrastructure, right? It has offchain indexing services that run to a network of nodes, but again not a blockchain. So now let's call those offchain services. So let's call them ABSS. Okay, ABSS, offchain services.

They could be work for automation, code processor, intent solvers, data labeling, inference verification, identity and key store. The scope is essentially unlimited, but they all have something in common. They are being built offchain. Let's take two step back and discuss what ABS is or the ads anatomy. Okay, ADS is offchain computing service that provides services to other protocols to other application to other networks and I will argue it's um those four pillars compiled of those four pillars it has offchain software offchain execution execution is being

done offchain but validation is done on chain ads is also a proof stake network because in order to opt in to validate the ads operator needs to stake a certain amount of tokens whether if it's Ethereum or the AVSA token they put their capital in line or in risk of of slashing and on the other hand are being entitled to reward and being incentivized to act truthfully is run by a network of operators. Those operators they run the ABS software simultaneously they execute simultaneously and they agree on the

result of the execution simultaneously. And lastly each ads has its own unique custom logic its own unique computational purpose. Now before setbacks and I want to discuss with you about our personal journey. We started off in the very early days of layer. We were part of I think the first five ads that were building on back then um has no tutorials guide books. We just follow the code base and started experimenting.

And we did this did this for 6 months. And during this time we basically understood all the complexities, bits and bites, the problems that other builders will have to face. And the thing is that instead of focusing on what we do best which is building our product, we were busy developing all the low-level info components, consensus on off services, peer-to-peer networking layer, peer authentication, discovery, notebook stopping, peer authentication, where the operator submit at the station, how does the protocol audit those at the station, distribute rewards, impose penalties and slashing.

But most importantly, how can we apply complex system design onto the ABS design at scale? Um, so after building a bunch of those components, we decided, okay, let's make it available for other builders and this is how we came up with the authentic stack. Very simply put, the authentic stack is ABS development framework. It's a development framework to build distributed system with offchain execution with offchain computation. Okay. All of the underlining um complex components around consensus and operators and stationation are all abstracted away. And the beauty here is

that we allow builders to focus on again what they do best which is building the product, building the service. Um, you can build in any code language, any use case. It's highly configurable and highly versatile and the amount of potential use as I mentioned before is unlimited from over to co- processors decentralized compute CPU clusters uh info for DeFi engines identity and key force exploit prevention system uh data labeling for AI AM finality financial systems AI system middleware intense solvers and data services and so on and

so forth. Top of our mind is to be agnostic to any type of use cases and as you can see our ecosystem is flourish. Again you can build anything. Uh we as a team we just love to think through design idea. We love to bounce ideas. We love to think through architectural design of ABSS. So if you want to build anything um would love to talk with you and thank you.