The Next-Gen Search Engine for Web3: Kaito - TOKEN2049 Singapore 2023

hello everyone can everyone hear me H everyone my name is Yu I'm the founder and CEO of Kito today I'm going to talk to you about the nextg search engine for web 3 we AI company we also web 3 company so right in the intersection of both a and web 3 and the problem was solving for is information access for the web 3 community and this is very crucial today because not only for the builders that are that are currently the web 3 space but also for the next billion new users going to be on boarded into the space so before we dive into what we what we're doing I want everyone to think about this question how did people live before Google information was extremely siloed right so if you Google is actually a company that's only 25 years old many people here are you know older than 25 years old I remember when I was little there's no Google back in the day people have to go to libraries magazines and then it becomes the internet so if you look at the revolution of information access there are four different errors the first phase is pre-web error whereby people needs to you know like I said like go check magazines go to library access information and very very difficult way and then we have the internet so information sted to come comes to the inter come online but at the same time this Still Remains extremely siloed The Searchers need to be expert in knowing where to find certain information then we have Google error now you don't have to remember all these sites you have one landing page that you need to remember and that can get you everywhere just now obviously since November last year chat gbt is launched now we're going through this transformation of chat gbt era people all of a sudden realize that you don't have to go through so many different search results now you can actually talk to a bot and then the bot were able to scan the internet or scan the database and to be able to answer your questions so that's a whole new different era of information retrieval but what what does that mean in the context of web three so if I think about these four different errors I want you to think about this question where are we today in terms of web 3 information access and I want you to read this pre preg gooogle error general information is online but remain Silo today in web 3 we have you know Twitter telegram Discord governance Forum Twitter space news research all these kind of very very siloed channels searches need need to identify the best websites for designed information today if you're going to you know go and find tvl you may need to go to defy Lama if you're going to go for you know protocol fees you may need to go to token terminal for example if you need to find let's say you know certain dashboards TV off for FR Tech you're going to go on Dune to be able to check this information so many decentralized um DS and put up a snapshot for votes but if if you want to see the rationale behind all these snapshots you need to go and check out their governance forums extremely siloed and you need to be the expert in knowing where to find certain information and this is making it extremely tough for any new users coming to the space the other day my cousin was talking to me hey I'm very interested in web 3 where can I get information I was like you're going to go on Twitter you're going to subscribe to or follow these a thousand people so then they know what's what's happening in the space my point is currently web 3 information is very siloed and fragmented and is very much resembles the preg Google era of what we went through in the past and what is that and the two main reasons why that is the state today the first reason is web 3 information lives on a very distinct infrastructure right think about how Google cross all the data is using a general crowler all the links on the internet and it backtracks all all of these links and be able to identify the relevant sites and all also come up with a page rank to tell you what's relevant but then web three information is silo by Nature it is and the re and the way to access all these information is also very different right so you need to set up a note need to go through all the blocks and all these chains are siloed as well you need to you know in integrate with ethereum then integrate with Avalanche all these chains are all kind of siloed so you need to go do this one by one is not and it's not compatible with the with the infrastructure that Google set up and hence why none of the um you know blockchain specific information it's accessible by Google and the second reason is decentralization and fragmentation and the way a lot of the web three ecosystem or projects work are also very different right so Unis swap team is you know remote um a lot of other project teams are also remote this is the way you know things are working right so make it out you know put put up a governance vote um but all the stuff like leading to the governance votes is actually decentralized in a way that Community can come up with a come up with a proposal these things happen in the governance forums by Nature is much more fragmented and siloed which makes the problem even worse and by because of these two issues the web 3 Community today suffer greatly from accessing proper information efficiently so we're still living in a preg Google error so what Kito does today and the problem like I said at the beginning the problem trying to solve is information access in the space where AI company we're leveraging so the call of the company is to build a solution that can really service the web three industry and our solution is we'll build a search engine from the ground up and then leverage large language model to emplify the user experience what does that mean so today the you know companies that are building on top of Google search engine right so the lags of perplexity is a is a very prominent unicorn in in s Valley but at the same time Google has come up their B which is also llm uh llm based search engine that's leveraging Google's own search engine and obviously new Bean is coming up with a new Alan based search engine based on the bean search engine but the but the fundamental information the the fundamental way they access information still through their own search engine but because of the problem that we stated at the start the traditional search SE Engine today doesn't index most of the crypto and web3 information and hence why we need to build a search engine ourselves to serve the web three community and the way we make the information or the way we make the user experience Great Is by leveraging a large language model to bring ourselves in the web three Community from the preg Google era to the Google era and then timate to the chat gbt ERA this is two orders of magnitude efficienc boost in terms of accessing information in the space so where are we today we found it a company in March 2022 so round about one and a half years ago received our C Dr financing sort of right about 12 months ago it was led by dragon fly sequ Jam Street and then late last year we launched the Alpha version earlier this year we launched the betza version and then today we're launching a public version this public version is not a fullon search engine yet I'll tell you more about this so this is a institutional phasing information service terminal it is meant to service all the people or the primarily institutions today because they spend days and nights going through information and research so we're targeting this Niche group first before we expand the mass Market to the search changing space I would like to show you a video in terms of what it [Applause] [Music] does [Music] [Applause] [Music] down oh [Music] [Applause] a [Music] great so as you can probably see so this is an Institutional phasing product and what it does is indexes so many different sources that researchers and and people in the space are constantly struggling to you know access and create and uh aggregate at the same time and cury so our institutional phasing protal can very much streamline the the workflow of all the researchers so where we're at today we built information service terminal for web 3 and this is what we launched today the next goal is next year we're going to be building a search engine for crypto that's going to be pH that's going to be retail facing it's going to be serving the mass Market that everyone's going to be free to use um a finally a web free search engine yourself and what's our end goal eventually we believe that the world's information is going going to be living onchain and offchain the likes of Google and Newan today they're going to index the all the information in the in the web era right so every anything that's on the internet but then we do need a search engine that index everything on chain and that is what we do and then eventually blockchain is not going to not going to be only hosting just transactions it's going to be hosting a lot of other you either you know articles that people publish decentral social a lot of people's you know tweets social information potentially images videos all these you know Publications all this kind of stuff it's going to be a Blog space for anything to be hosted on that we do need a search engine that is able to index all of this information and that is what we serve and lastly so this is this is our contact information we also have a booth at token 20 20149 so if you're interested feel free to come by happy happy to have any conversation with you and any collaborations that you want to that we you want to you want to build together we're very happy to take and we have one more minute so happy to take any questions you may have um are you directly integrating with like Twitter API and the Discord API like how do you how do you pull the information is is a mix of both so we we are having official um a partnership with them and also for Discord you are we essentially only scraping all the public service so we're not touching any of the private service yeah so all any but all the all the project public uh all all the project Discord service all public so it's yes yeah only index the public ones yeah yeah that's right yeah um hello you here uh you have you said that uh as you said that the on data is unstructured how do you Index this do you have a layer of interpretation how you deal with it so it's a very good question and then the way we think about it is obviously a ton of companies building the onchain space right so we're standing on the Giants of the well shoulders of all these Giants right so they are you know nanson Dune def Lama token Terminal A lot of onchain companies have done tremendous work in the space cing very noisy blockchain data so the first step for us is to be able to identify and query all these pre-processed information layer data layers so for example if people are asking questions such as hey how do I find a tvl for frch we want to pinpoint them and direct them to the Dune dashboard that specifically solves that obviously we're aggregating all the other sources as well so like token terminal and other other places so that we become a search engine that people can rely on in terms of finding the information and people don't have to necessarily remember hey I do need to go to that side for that specific piece of information hey maybe you know Dune doesn't have that maybe I need to go to flip side Etc so that's that's our approach I think I probably r out of time can we take one more oh okay very happy to uh I know there are more questions so free free to come out the booth I happy to take the conversation further thank you