Web3 AI & The Graph by Charchit Kumawat

Transcript

GM okay so hello ladies and gentlemen and myself charit I'm the community lead India from the graph so today we are not going to dive into too much technical about the AI advancement in web 3 but rather we will be focusing on more towards the application and use use case side of it so the main so today we are going to talk about three major points web3 data Ai and the graph so like in graph we generally don't talk about market valuations pricing but still it is one of the most important metric to know about the web3 adoption so last year if you see uh like web3 adoption was valued around 23.3

billion but in 2025 it's expected to grow uh towards 39.7 billion USD so like if you see like if if like the web3 is getting adopt uh getting adoption so as the query volume within the graph is increasing so from the 2022 quarter 1 to uh like last year quarter 4 we have seen approx 121 cagr growth so that's pretty much of how fast the web 3 with the graph is going towards yeah and now the main question comes why Ai and web 3 yeah so before going further let's understand slightly

basic about web3 data so the property of web3 is uh web3 data is like Distributing data infrastructure so what does it bring to table redu sensorship improved accountability and uh immutability and then the question comes what AI can do so AI can bring smart insights personalized experience and then very basic uh like let's understand what it can do in web3 so like imagine having your llm trained on a web3 data like having like since the web3 data has the purest uh purest property of having a so track and traceability and

verifiability and you train your llm on top of that web3 data you can take much more informed decisions and with the informed decisions you can reduce the security and eventually uh like uh it will also help you in resource optimization and then the uh like the third part comes what uh the graph does in this segment so in simpler words graph makes it easy to fetch your blockchain data and ensuring verifiability openness and permissionless leer So like um like the main question comes what's the need of the graph right so like in the past four or five years there were many startups or many media houses who were dependent

on Faceook Facebook page Google WhatsApp and uh like different kind of social media so like like when when the political environment of the world becomes more volatile it is not very much new that these are the startups or these are the companies affected by most so the that's why the graph is focusing on to enabling full stack de uh decentralization which will ensure the application or your startup won't disappear within one day and like the best part about the graph which I like personally is our solution is governed by the graph Network where developer can create their open AP subgraph the indexer can use the fire hose and

substream is like a newer tech building by the simor uh like streaming fast uh AI team working on the graph and then we do have the substream power subgraph and like let's talk about the graph Network so graph network is supported by Major Le three players indexers curator and delegator and so for example if a developer like you and me deploys a subgraph over a smart contract so like on that stage your subgraph is empty so it so the curator needs to understand the business value of that particular subgraph so like once he understand the business value he will curate the subgraph with the GRT tokens

and then the indexer will see oh the curator has signaled the the subgraph with the GRT so probably it will be having more business value so indexer will come on board will index the data from the blockchain to the subgraph and then again there is the delegator who will Who is responsible to support the indexer by delegating some uh like grt2 indexers and for the like once the process is completed the developer will pay the query fees in in GRT and now it it is like now we have also started supporting credit cards so the de will just need to fetch the data from the gra uh from the subgraph using simple uh graph ql and now we have understood the basic

three points web3 data Ai and the graph now let's understand more about the application side of it in going in more depth towards it so like as I mentioned like the developer need to pay the query fees in GRT so like still uh the query fees is still slightly high right now plus and uh there is inconsistency in the query uh query fees so here the semiotic Labs comes in it's a codev team working on the graph especially in the AI segment so they developed a domain specific language name Agora so Agora helps to categorize the query uh query volume with uh like

with the proper shape and that's how they are uh planning to uh optimize the query fees and then the AI part of Agora comes so the first major problem of Agora is it it needed to be manually installed on a particular subgraph and like so the auto aora is AI enabled and it is personal like it could be p uh we can use autoa for like for the personalized subgraph and it could be like uh more more effective and uh better towards using on the other subgraph and now like let's dive more towards the llm trained on the verifiable data so like having the property of track and

traceability in the blockchain helps you to get uh like pure data from the from the uh web 3 and imagine like if you like again train your data over the llm uh llm models of yours like so like it could also enable new applications within web3 so in the current world like in most of the trading uh like financial markets Traders are using python Bots to trade and there were many risk of manipulations and fake news unnecessary negativity so the llm trend on the verifiable data can improve the decision making and help you to take more

informed decision within D not only that in Dow you you will uh you you will start like you can have the ability to uh track the fraudlent activities within the Dow and you can easily like uh curb the uh curve the bad partner within your ecosystem not only that uh the llm trained on the verifiable data can also help uh in the developer tooling specifically in analytics so once uh like you have your AI powered analytics over the purest form of data then a lot more applications and lot more business uh business value a local customer or like a local business or any particular

startup can start having and not only that the graph is also working towards like uh towards uh helping the developer to select uh select the uh select the proper subgraph for their business business value so for example if you if you want to make a market maker there are plenty of subgraphs within the graph Network more than thousand subgraphs are already there but still it will be difficult for a new developer to select the right subgraph for for their main purpose so the AI will simply help him to select the right subgraph and his query will match to the right subgraph with the help of a so that's where the

graph is working and now let's talk about the community so since I Le the graph India Community like uh like I I would like to share some glimpse of it so like U here in the graph India Community our focus is to scale the uh ecosystem of developers and uh like our Moto is to empower them to so that they make industrial gate subgraph and like a small context of web3 in India is like a lot of projects are coming from polygon uh like coming from India like polygon bonom instab so it's not very far that we uh cannot expect a industrial great subgraph from a community so in the next

uh coming year like it will not be wrong to expect that we will be having more industrial grade subgraph from this community so that's all like if you have any questions feel free to ask and like in the meantime you can connect me over Twitter and telegram like I I am very responsive and if you need any help from the graph for your sub uh particularly your subgraph or any uh data service you are looking for feel free to connect and yes uh like if anyone has having any question feel free to ask cool so thank you everyone and yeah

take care make sure keep building uh and drink lot of water yeah