The Rise of AI in Web3 Development UX by Stephane Tetsing | Devcon SEA

[Music] yeah hello everyone my name is Stefan tessing and I work with the remix team and my talk today will be about the rise of AI in U web predevelopment area I will be much more on the application site site and yeah regarding at the uh at what uh the AI can actually bring to the web environment these are actually the main application fields that AI can bring a lot to first of all Reas the user experience where we can incredibly um enhance uh how the user actually perceived the platform and also the decentralized governance that AI can actually have to much more incentive incentivize this uh this term also the creation of digital assets also deals and we cannot also we cannot speak about AI in web Tre without talking about fraud detection and also the security aspect so actually enhancing the user experience is pretty much easy here we have an example of the remix platform where we have code completion code explaining those are pretty much low hanging fruits for those who knows how llm works and yeah this actually also already provides a lot of yeah a lot of input for the user a lot of value and next stage of uh providing uh value for user is by having um automated uh automated chat boats like also AI support which are integrated in the platform here again a remix example where the AI chatboard is aware of what the user is doing is always one step ahead in order to kind of provide the user a much more better experience and so forth yeah also the AI can be used for creating um personalized content user feeds creating also AI board that will be serve as Oracle for contract and also provide investment recommendation in create digital ads we as we have already seen with nfts and uh automatically generate metadata for Market visibilities on different platform without yeah sharing the asset regarding at the security uh aspect of how AI can actually improve how we develop uh a web threee application uh the it can actually provide a much more secure development mechanism just think about it when you're about to deploy a contract and there's a AI around it that kind of that kind of pass to uh the contract and for vulnerabilities and check pretty much everything what goes around it and also trans transaction there's AI maybe sitting somewhere watching the blockchain out real time and kind of flagging um fraudulent transactions and so on so the AI sitting somewhere and kind of doing um behavioral analysis and profiling everything and yeah flag bad stuff okay the last thing I will be talking about is about agents uh agents is like like a software layer on top of yeah AI model which actually can perform autonomous uh s and having interaction with AI in order to get out out in order to provide Bas values and those agent are actually able of kind of doing uh vulnerability scans uh help into improving the decentralized governance and also regarding the uh decentralized Finance we can automate uh that U that aspect of the web Tree in environment by kind of providing uh best trading pets best price best algorithms for for that and also the smart contract automation that the agent can actually help with and yeah I think that was pretty much my talk I am very much open to questions yes thank you very much anyone with question I see the gentleman in the very back I don't think I can throw that far thank you uh hello hello so I had basically two questions my first question would be regarding the remix AI so I actually you utilize remix a lot um for contract development so what I wanted to ask was like the anwers That Remix AI provides how how does remix handles like hallucination over it the second is like so remix remix AI on the back end uses like llm so is that llm open sourced and what are the basically things on which it is trained at like Are there specific contracts on which it is trained at uh first of all regarding the hallucination uh I think it is a problem that you cannot actually avoid with the noway stage of development but we aim to kind of reduce that uh uh by utilizing uh uh state-of-the-art models and to be much precise regarding also the second question that will be the open the Llama model we use the latest llama model which are available on hugan face and yeah they are deployed on our backends and very soon they will be able you will be able as a private person to run it locally on yeah at your own risk as I will say another gentleman over there yes uh sorry I'm thank you very much very nice box thank you uh I hear here a lot of things that a lot of people are talking like how we can do better contracts how can we do better chains and uh I feel that we're talking too much about uh the core layer like how to make the things run and uh we're not talking enough how to build a new things over it like web apps Journal apps do you things that AI can uh improve here too or like will it be more bring bringing more to the table for uh client part and uh high level Builder part or it will be bringing more to the table for lowlevel like blockchain and contracts I think it would be much more on a high level perspective I that I will bring that value uh okay go it thanks yeah I think we have time for one last question if there's any one last question over there there's a gentleman there you yes amazing hi hi my name is Morris I'm working on 42. money we're trying to generate front ends for smart contract specifically um what is your vision on web 3 uh generation of UI uh we see a lot of big players in web 2 but in web 3 there's not really uh competition yet I mean my vision or my the way I say it is that the future there is very is very much bright we are right now at the beginning of a of the era of mean AI casing the web pre environment and I think it is pretty much open you can do pretty much everything you you would like to do yeah okay thanks okay