Surfacing Signal from Noise: Data Insights from Research Powerhouses - TOKEN2049 Singapore 2024

well we're going to have a little bit of fun uh on this panel I'm Samantha I'm the founder of Yap Global we're Pier agency helping web three and crypto companies tell their stories but numbers and data actually speak louder than words and so I'm very privileged to have the research powerhouses in the industry we've got Alex from nanson Yo from kaito Ai and Matt's from Dune but let's start with the story actually so Alex do you want to begin with a a short little story and you guys can share a little bit more about what you guys have seen this year absolutely great to be here on a great panel uh I want to share first because MZ and I used to work together a long time ago and I remember in 2017 we were sitting in Barcelona at some bar talking about how we need to start working in crypto and here we are many years later um we've created the two leading onchain data analytics companies so it's great to be in a panel with matut and of course with you who uh we spend a lot of time together as well here in Singapore so yeah but yeah I run nsen which is uh onchain analytics for investors our mission is literally to surface the signal and create winners so yeah happy to be on the panel Matt Ste go next yeah I just want to jump in here because the the way I got into crypto was Alex convincing me to invest in an Ico so it's it's quite fitting that nonon uh or that Alex went out to start nonon which is investor Focus clearly but yeah um I'm matz I'm the co-founder and CTO of Dune started Dune in 2018 our focus is to help blockchains protocol teams and users to understand what's going on on chain um from a community angle so if you log on into June your dropped into a product with half a million analysts and uh yeah that's us yeah yeah I didn't I didn't have the privilege to be working with Alexa mats directly but I did come into the space using the products all day long so definitely kudos to them my name is Yu I buil um kto which effectively is a search engine for the web 3 space which is now the leading information service terminal for for the space um at this stage we help more than 300 companies half of those are kind of a funds um basically liquid funds and VCS institutions market makers Etc and the other half product teams such as ethereum Foundation salana barain I these effectively project teams and funds surfacing signal from unstructured tax information many so that's kind of the key difference versus maybe some of the data Gres on the stage we don't track onchain transactions per se but actually we track onchain and offchain unstructured tax information thinking about Google for web 3 yeah we were just chatting backstage about how impressive it is being here at token 249 with 20,000 attendees and all the beautiful booths that you see outside but it's noisy right I mean how do we filter out you know what's the leading crypto protocol what's doing really well and so really you're you're on stage here you're looking at the go-to data analysis you know research powerhouses so just to begin and this is for uh everyone and anyone to jump in on could you give us a recap of of the year um the past year how has crypto analytics evolved um and yeah maybe we just begin with that yeah happy to start with some thoughts I think the thing that's quite unique now and also I would say last year compared to a couple years ago is that uh we live in a multi-chain uh and multi-layered uh world and that means it's not just more complicated for us to to do our work but it's also much more difficult for people to stay on top of everything that happens on all of these different chains on l2s even l3s and so on so I would say that's a big Trend I don't think that's going to change I think we're going to continue to see lots of new chains and l2s come to Market uh and and in a way it becomes even more important to curate all this information so that you can stay on top of everything that happens because you don't really want to have siloed views on each single chain if you're tracking your portfolio you want to see that across all of the chains if you're looking at where the smart money is uh investing you want to see that in the context of all of these different chains so I think I think that's a really big Trend and I'm sure something that all of us here uh relate to yeah that's interesting there the different multi- chains and how do you really keep track of it all uh how does dune Dune analytics do it yeah um so Dune obviously has like a uncurated look at everything that's going on today I think we have 75 blockchains on Dune and it's um you know pure chaos I would say like there's so much data and the trend for this year I think is that um there's so much available block space and so so many new applications that are being deployed that there's there's actually meaningful insanely meaningful activity that's going on and sifting through all that data is challenging but even more important than before yeah oh that feel greate with that actually I mean it's Alex obviously talked about the chain um from our experience even if we look at kind of the the the world outside of crypto you're actually seeing more fragmentation in terms of dates think about nowadays you know once llm comes around you actually have all these social media platforms shutting on access right nowadays if when you use Google you can't actually tape into Reddit you can't type into Twitter you can't type into Discord Etc there's actually more fragmentation and multichain is kind of one you know angle and aspect of that as well so I think nowadays actually we're facing a unique situation where internet is more close than ever dat is actually more fragmented than before so it's this is actually more relevant and more important to actually be able to derive insights from you know just such a um such a fragmentation um structure and speaking of our experience maybe slightly different from um you know obviously ding at nson being around for a long time K actually just got launched in the beginning of the year as a younger company uh we obviously being seen you know pretty pretty good growth U adoption from both kind of the fun side as well as from the project side I think a lot of the stuff that we are doing is is trying to kind of standardize um in terms of how people are thinking about the industry thinking about approaching a different stuff like for example from the font side we're now having index for social cerament we're having index for mind share we're having you know narrative uh mind share trackers is for people to really understand without actually given without actually going to the very specifics of all the people um like all the Twitter timeline and how people are been talking about stuff but actually very easy to synthesize in terms of hey like mcoin uh was kind of a major narrative kind of going to March and then now it's kind of Defi and some of the fundamentals coming coming around and train abstraction all that kind of stuff we're able to kind of abstract way all the specifics of what people to been talking about actually um you know surfacing insights from that perspective obviously from a project te perspective people are now using kol for evaluating their campaign um you know tracking the Mind share understanding where some of the growth Dynamics whether they're under o under over index in certain regions all that kind of stuff so a lot of what we do is really coming from complete Bank piece of paper and trying to Define how do we approach basically measuring some of the key kpis and thinking about some of the mega Trends how you're doing within those Mega Trends yeah so in summary there as um Matt you were saying there's 75 chains to track there's so much to keep up with in crypto uh but there needs to be a filtering and really nanson kaoi and dune are doing that for Traders and for teams so yeah that's that's amazing work you guys do and I think it's really valuable we're here in Singapore uh I think a significant portion of the crypto industry is here in Asia which is super exciting uh I know Alex and Yoo you live in Singapore so are there any interesting trends that you see regionally you know uh yeah what do you see compared to like in the Asia Pacific region compared to what you see with crypto activity in like us Europe and Mia are there any interesting insights you can share yeah I mean I think uh Singapore is kind of rising as one of the main crypto hubs in the world um you you you probably conclude that I think New York City is the number one crypto Hub globally and Singapore is probably number two um and then Asia is actually incredibly diverse when it comes to crypto like if you go to Korea it's a completely different world uh Korea's one of the most um active markets for for training and some of the largest exchanges have Korea as their top Market Japan also of course had kind of a um somewhat Rocky history with crypto where they they over regulated the industry and sort of strangle it but it seems like it's coming back now uh and then of course you have so many Chinese investors and um Founders in crypto who live actually all over the world many of them are here in Singapore uh so so I I I'm personally really bullish on Asia overall and if we zoom in on southeast Asia there's a really interesting ecosystem here around web 3 gaming for example where you have uh Ronin that's actually technically at least they have an entity in Singapore but of course a lot of the team is in Vietnam and a lot of their players are in the Philippines for example so you have I think I think Asia is super interesting overall for crypto and many of the different countries play different roles somehow so yeah I'm really bullish on the future of crypto in in Asia I think that over time Asia is going to become the most important region uh for crypto but of course you know today I I think you you you'd probably conclude that the US and specifically New York City is number one right here to hear I know that Alex obviously always been a a huge Singapore bull I I enjoy reading all your tweets uh around that I actually I I over time I grew more and more bullish on Singapore and actually I mean uniquely I think obviously I agree with you like New York is number one in terms of just being a crypto Hub I think what's unique about Singapore is is actually here we have a very strong pool of a very very actually very balanced pool of talents so you have obviously like a lot of diversity in terms of access to all the Asian countries but also at the same time even for us we're able to tap into some of the machine learning Engineers locally which we wouldn't otherwise have access to any in New York CU a lot of the machine learning um talents actually on the west coast but actually here we also have a bunch of other talents that BD talents we have you know data scientist talents all that kind of stuff so I think it's actually a very very good mix of talents that we can tap into locally Matt what's your impression I mean you and I live in Europe but it' be good to hear what your impression is coming here and and seeing what's happening I must think that this week is something special because it's absolutely insane at token 2049 with 20,000 people I don't know how many boots there are here but I've seen more companies more chains that I've never heard of in my life this week than before so it's pretty intense I live in sug Switzerland which in many ways are like the Cradle of crypto I think because of the extremely forward leaning uh regulatory agencies that essentially protected utility tokens Etc so my office is 200 met from the ethereum foundation for instance and so it's definitely very different from Singapore the scene is very small um so yeah it's great to be here uh amazing energy i' say maybe there'll be more protocols and um you know chains to track on your an analysis platform but now let's move on to narratives uh crypto is one of those Industries where every year there'll be something new that's trending you guys were talking about back in with icos then it became stos then defi summer then nfts so let's go a little bit deeper this year and over the past year we've seen terms like chain abstraction decentralized Social Web three gaming dpin crypto AI which we'll delve into AI in a little bit and then defi just which of these narratives do you think you know has the biggest chance of growing in the next year or are there any yeah insightful you can say about what you you've seen in in terms of the reports you've tracked lately anyone can jump in yeah depends a little bit on the time scale that you have in mind I think the one that has the most potential um on a long enough time scale is definitely rwa real world assets right because it sort of brings in external forms of value and yield into the crypto ecosystem right I think even vitalic had a tweet about this um where he he talked about how the yield should come from some external place right so if we think about stable coins as the first rwa uh with product Market fit although many people don't think of it as an rwa because dollars is kind of an abstract thing but I think it it is the first example of that and then you have tokenized t- bills and money markets which have been growing very steadily and you have lots of different uh protocols and projects doing this um now you're seeing things like like tokenized uh gpus you know and you're seeing of course um eventually I think you'll get to the more sort of physical you know real world things like property of course real estate eventually is going to come onchain and if you think about just the value of all these real world assets both in uh the form of collateral that you can borrow against if you have you know a property as an nft and you can borrow against that I think that's a huge uh growth driver for crypto but you kind of have to do this sort of asset by asset and you know my My Hope and sort of the future I want to create is a world where every asset is tokenized that's going to take a long time and uh I love the name of this conference token 249 because it gives you that you know future perspective uh but I would say like rwa I think is one of the things is just steadily going to grow and we're g to sort of overcome each hurdle uh at a time and you're going to get new assets come on chain step by step and that's going to grow the whole uh ecosystem you know I think one way to think about the craziness of meme coins and nfts and so on is that these are somewhat speculative use cases in the short term but they actually help create the infrastructure that can be used for rwas for example so if you have an nftf protocol where you can collateralize an nft and borrow against it today that might be a H penguin but you know in the 2030s you might be doing this with an actual house right and so to me RW is the one that I think is the most compelling it's also the hardest one in some ways but yeah I'd love to hear Yo's thoughts as well because you guys track like narratives very closely so maybe you have some thoughts too I mean definitely Alex in terms of uh you know obviously our rwa has been a kind of a major major Trend um over the past 12 to for month um as as he as he pointed out we actually track all the narrative um in the space and the way the way we we track it is is basically just by measuring the the social mentions of these terms from our tracker the number one narrative has been AI very very dominant and number two and number three has been mean coin um as well as defi they've kind of been rotating across but there also a lot of the the emerging narratives that that really gaining a lot of the Mind share chain abstraction being one one um dping being one these are seeing really really significant growth either from adoption perspective or simply from a from a mind perspective what we are also seeing from a different perspective is there there actually seems to be a lot more focused on fundamentals and Common Sense at this point of the cycle a lot of the times that people are asking the hard question in terms of where does this need need to be on the blockchain what does this really kind of do to the kind of uh the overall adoption we see more and more people asking questions such as this which also kind of um resonates with the recent trend of you know the the so-called defi uh Renaissance because people are really kind of going back to some of the real adoptions and and thinking what are some of the real use cases that that people actually adopting at this stage of the cycle so I think that's a that's a good Trend that people are applying common sense to all of the either from a fundraising perspective or from a technology you know building perspective so I think that's you know something that we're seeing as well yeah yeah U I'd love to jump in here because I think the obvious trend to me is protocol teams turning into chain teams so essentially a new web three native business model has emerged with this sequencer fees so um we work with a lot of protocol teams and a lot of them are at the point or where they're considering they have or they are going to create their own rollup and I think this is super unique because now the blockchain is actually a piece of technology that you could put to use to your product it's not just you know a global database that you uh program against but it's something that you can purposefully use in your product and you can you can have composability with all other apps and say the ethereum ecosystem but tune your sort of Economic Security and I think that's a real Trend that's going on I mean we entered the panel talking about multi-chain I think that's going to continue we're going to see more chains more protocols being embedded natively into rollups and um yeah I'm excited for that yeah going one step deeper as a followup to that I think kaoi recently tracked that chain abstraction narrative is really picking up uh C can someone elaborate a little bit more of of where you see that going because I think yeah from a c abstraction to chain abstraction that's on a more technical layer that seems to be something that a lot of protocol teams developers are talking about lately I I can CH in here obviously I think it's it's not secret that we're we're now seeing a lot more chains and also seeing um you know one of the major hurdles for user experience is nobody wants to be um you know constantly bridging across so many different chains and um to be able to you know basically set up different wallets for different chains U liquidity fragmentation all that kind of stuff it's a whole wave of really you know different kind of developer activities and kind of investment investment dollar really going into that field to really trying to try to build a better user experience for the end consumers yeah the most magical onchain experience I've had this year was on Sora where I was going to me an nft on I think it was the Sora network but I have never bridged to the Sora network but their app allows you to pay on any chain that they support so I think I actually paid like usdc on Bas and I received the nft on Sora chain and this is powered by relay protocol it's the super magical sort of uh bridging product but essentially it's chain abstraction right um Sora doesn't care where my money is um and I don't necessar neily care which chain the nft is on and that was super magical I I think we're going to be seeing more of that hopefully and and Mak it more and crypto being more usable uh for people I think vitalic spoke about that yesterday about crypto being more usable for everyone youo you you brought up that AI is definitely a big Trend that you're seeing um so with the rise of AI machine learning um and especially in crypto analytics I mean it's in your name well nanson doai is your your site and then Kao AI it' be good to hear what you guys are doing with AI and and it's the role that it's playing with uh crypto analysis yeah we I'm sure all of us could spend hours talking about this topic actually um I I mean I think every company has to obviously have an AI strategy every software company for sure and probably every other company uh the way we've approached that nson is to think about it in two main ways one is how do you let AI power the product that you're building and the other one is how do you use AI for organizational productivity and so on the productivity side uh this year we've spent on average about 5 to 6,000 hours per quarter collectively working with existing AI tools so this can be as you know as mundane as a research analyst working with chat gbt to write the research piece faster or an engineer using Cur to write better code faster um or myself when I write my CEO updates I'll go for a walk and I have chaty BT voice on and I'll just you know bounce back and forth with chat BT to write the actual message to the team um so this is the productivity part and and I think that's kind of universal to any company you have to think about how to use Ai and make your team AI native uh on the product side which is more specific to what we do there basically three pillars uh for nonon the first one is we use AI to build a stronger data mode right and many people know nonon for our ability to label addresses right and now a lot of that labeling is is fully driven by AI actually and so this is pretty amazing just from like a cost perspective because the way we've done attribution or labeling of addresses we've always used the combination of of man and machine but now the lines are kind of blurring a little bit and if you think about the cost structure of having an individual finding out what is this address and who's the entity behind it the time that goes into it you can associate an economic cost to that and when you do it with AI you can just do it literally with a 100 or THX lower cost which means you can scale it up and label just way more addresses so this is about the data mode part the second part is to uh allow for fast and fluid interfaces and this is the the more kind of futuristic part of of how to use Ai and the one that's going to take a long time to fully realize uh the way I think about it is if you've seen the movie Blade Runner the original one there's this scene where Rick Deckard uh played by Harrison Ford bounces back and forth with a with a TV screen and he's zooming in panning out and giving voice instructions to uh to the the screen to figure out what's actually in the photo I always thought of that as kind of the the voice and vision combination which I think is super relevant for all the stuff that we do on analytics right because if you're doing analytics and decision making you want to have some kind of visual support for your decision- making it just gives you more comfort and I think the voice and the natural language interface is going to be really powerful so we're trying to Trend in that direction obviously today you can't really speak with nonon but like that's where we want to get to over time and the third part is personal ization uh which is super under underutilized in crypto like in web 2 every product you used is personalized but in crypto somehow you have very little personalization so for us when we help investors we want to take your portfolio your trading history and actually help you surface new opportunities as well as to give you input on what you might want to watch out for based on your trading history and this of course is uh something you have to do with AI at to make it work at scale so yeah building a stronger data mode building fast and fluid interfaces and personalizing the whole user experience I agree with the you know all the all the points that Alex mentioned I actually think that AI the application of AI within crypto is actually very very important actually goes really well because crypto you do have all the onchain transactions but actually without being able to kind of properly abstract way or be able to understand contextualize it is is very difficult to understand similarly from the unstructured text perspective you have such fragmentation and so much noise CPS is actually notoriously known as just a lot of scammers and a lot of spams in the space how do you actually apply AI technology to actually basically future out all the noise and you know surface or the signal is is a huge undertaking so as as Kito AI you know funny background actually part of the reason why we were called Kito is because there's AI in that so there always been kind of AI company building a search engine in the space search engine obviously um is an AI product but actually with large langage model the whole way of how we're doing things has completely changed and I kind of echo with what Alex mentioned in terms of maybe the product side of things we consider as exter external user interface but also kind of the internal um kind of things that effectively how we're doing things and both of these have completely fundamentally changed just maybe giving some examples here internally back in the day if you kind of maybe just like using Google right the expectation of once you search for certain stuff the expectation is within the three the first three Blue Links you need to get to precisely what you need to get to otherwise they think the search results very bad but nowadays with large language model as long as within the top 30 results you have one that's really precise can get you what you want you can actually use large langage models to go through all of this and then trying to semantically understand this and get you the answer so the whole Precision rord problem has changed or similarly for example for us back in the day you actually need tremendous amount of user data to be able to train your machine learning model nowadays you can actually use basically agents and Ai and machine learning models to actually just constantly re-evaluate itself so for example the way we're doing search nowadays is really just you know the the the agent will essentially change the small parameter and then evaluate both results and compare which one is better and then on the back of that we can just constantly iterate so the whole whole way of how we are how we're building the search changing how we're improving and constantly iterating on the back of that has has completely changed and then kind of coming towards the the the external user uh interface perspective I think a lot of the people um the moment they they think about large language model you think about kind of the the chat B interface but actually it's it's it's much more than that right so it's all the output uh the Precision of all the output that back in the day even like nowadays for example you have a lot of the crypto projects names are story movement blast these are very very generic terms how do you do proper social listening back in the day there was no very good solution actually involves a lot of the manual and human labeling nowadays you can actually basically use context based tagging I'm understand semantically what is tweet is talking about all that kind of stuff that's that's a game changer in terms of user experience but might not be kind of a chat interface although even on just on that front we just recently launched a co-pilot which is is effectively a chat gbt interface but with realtime crypto knowledge and information and then we're working towards um launching our fully autonomous research agent step by step we essentially de democratizing how people are interacting with information and these are the stuff that just you know wasn't possible before and now it's it's completely transforming the way um things have been done that's super interesting as someone in crypto PR and we also have to keep up with the trends and narratives I find that very insightful Matt what's June doing with AI I was going to say you've opened a Pandora box there because yeah so much AI there's so many dimensions to this question because like you have you go through the boots on level four and five and you have probably I don't know like a 100 crypto xai companies doing anything from decentralized training decentralized compute verifiable inference all kinds of stuff I saw this epic meme on Twitter which was like uh crypto uh saying that U crypto is necessary for AI agents to transact in the future and then the AI people is like I don't think about you at all yeah like I think we're definitely trying to um combine that I'm not sure AI is thinking about us but on like like we're data companies we're going to eat anything Ai and I actually think generative AI is most applicable in data companies like uh fundamentally it's information retri and yeah we've put it front and center in our product we generate SQL for our users given a prompt we explain dashboards explain queries we let you Fork queries using a prompt um like 99.95% of users uh automatically fix errors in their queries using our uh debug thing so it's like it's really transformative from a user experience perspective like both the chat B interfaces but also the fact that you know it could operate over TX so efficiently and I'm sure like 100% of your engineers are using co-pilot or cursor or something like that we are too and so it's just really like revolutionizing everything I think yeah um as we wrap up um we want to let let's talk to the audience in the room here the the the the trader the retail investor how can uh your research um firms uh support the the the average Trader in in the room and and yeah H H what advice would you have uh to them navigating the market right now yeah I mean nonon has kind of overtime uh crystallized its focus on the investor right in the beginning you had sort of journalists analysts and soone using our product but we've just doubled down on the investor um segment and the investing use case so with nson you can track your portfolio across 50 plus different chains you can discover tokens that you otherwise would not hear about uh through our product and then you can perform due diligence on those tokens you know within the same uh product but in addition many people use it to discover addresses and so one thing that a lot of our users do is that they find who are the most profitable addresses to to follow or to learn from and I always thought this is interesting about blockchains because they're inherently transparent and so you can see among you know the hundreds of millions of addresses that exist what are the few thousands or few hundreds that you actually want to follow and see how much money they're making which tokens they're buying and so on you don't necessarily have to copy train them blindly you could do that uh but just learning from the Masters is something you can actually do because of the transparency of the blockchain so uh this is something I I think is super interesting kind of an emergent use case that has come up after we started on um but yeah so I I would say generally the good thing about our product is you get this kind of vertical integrated product where you can do all the tracking of your portfolio everything on chain and also do the sort of alpha generation uh as well as the do diligence and uh yeah copy trading is one of the things you can do uh either directly or just to take inspiration and learn from the ones that have the highest uh p&l and the highest win rate that's the most exciting thing about defi it's so transparent compared to traditional Finance where you don't really you're not if you're not in the room you don't really know what's going on and how how other Traders are doing things so that's exciting and it's cool to see that there's platforms where you can track these things um you who yeah so I so as as K told we are a little more kind of general purpose as a search engine fun F actually I I came from tral Finance I I was a Trad at s for about five years so initially the reason why I buil Kito was because I was struggling with my research in the space because that none of the processes I had back in tral finan is working here nobody files the SEC filing and people are just tweeting on their personal Twitter account and sending announcements on the on the Discord so I had a lot of a struggle kind of doing my own research and hence why I build a search engine essentially helping funds and Traders and VCS essentially be able to do proper research manage the portfolio Etc over time unintended consequence uh unintended use case actually we realize a lot of the project teams actually need need this because people want to do you know proper social listening which is very very tough to do on all the different social media um people want to understand how they're doing on the social mind share how they're being perceived hence why all the sentiment trackers are very important for them people want to understand how they can grow whether they over on the index in certain regions um if they want to Target certain cars can I can I have organic way of targeting certain KS right rather than just generically reaching out to uh a group of people who uh want to be helping with your campaign so all of that is effectively some of the derivative use cases that we're now we're now serving I'm sure there's a lot more that we're coming up with and I mean so I either basically either your project team your Builder your investor your Trader your are Exchange Market maker you can essentially come and come and use Kito for surfacing insights from all the noises yeah it's uh interesting you say Twitter or X I like to say that X is like the public Bloomberg chat room of the crypto industry and that's why we need you know platforms like Kido I know a seat for a Bloomberg terminals like 20,000 and these guys are charging a fraction of the price so do consider uh Matt uh yeah any advice for um how you know the retail Trader can gather some insights from Dune and yeah what what advice Happ them June is not focused on the investor at all June is a product for the Builder but that actually makes it really good for the investor because uh Paul Graham the founder of YC says something like if you want to know what the next big thing is look at the the builders and the developers what are they doing on their weekends Etc and so over time dunit become sort of like a front page of crypto and you go to Dune and you see what's trending and that gives you a good feel I guess of um what's actually happening right now and so we're not really targeting that Market at all uh we leave that to our friends over here but um yeah like you look at Dune and you get a feel for what new products are coming out what are the airdrops what are the you know things that are happening on chain there we go you guys should check out nanson Kao Ai and June thank you so much Alex Yoo and Matt for your insights please give them a round of applause