Navigating Blockchain and Web3 Data

Transcript

welcome to the panel today on blockchain and web3 data I'm a moderating I'm a moderator today my name is YooHoo I'm the foundation of kaito kaito is building an AI power search engine for web3 well part of the one of the newer comments in the space trying to hoping to contribute to the space today we have five ogs on the panel so super excited to get the ink size ahead of everyone so before we start let me give the floor to the panelists today so maybe we start from Alex and just give a quick introduction in terms of what you do or your role and what your product does absolutely hello everyone my name is Alex tanovic I am the CEO and one of

three co-founders at nonsen nonsen is an on-chain analytics product that crypto investors use to better understand what happens on chain we also help crypto teams chains blockchains blockchains exchanges protocols better understand their users their competitors and the ecosystem around happy to be on the pan yeah I'm uh Frederick CEO and co-founder of Dune we make it very easy to query blockchain data and then visualize it share it with the community we have a large community of people coming together to understand what's happening on the on the blockchain excited to be here

hey everyone I'm Rafael I'm co-founder at last note we're a launching Market intelligence firm which means we provide Data Insights and tools to professional and institutional investors we've been around for five years now and yeah are proud to serve hundreds of institutional investors today from hedge funds asset managers family offices and everything in between cool I'm Mark graniger I'm the CEO and co-founder of genelysis it's analysis again data analytics looking at all the data on all the blockchains out there we have slightly other use cases than than

the rest of the panel here today so we use our data for assisting as investigations for governments we work with 50 governments in the world around 200 government agencies around the globe are customers of ours and on top of that all of the crypto business you should know out there a thousand customers in the crypto space we Empower them to do compliance and actually follow the regulations that's out there and enabling the industry to flourish hi uh good afternoon everyone my name is Matthew and I'm the CEO and co-founder at etherscan our mission is the scan is to bring Equitable human readable blockchain data

to the masses and the way we do it is by building block explorers blog Scholars are basic tools that allow you to understand what's happening in blockchain and to also interact with it in different unique ways thank you it's great I I know that Raphael you just mentioned that uh it's like last night has obviously been in space for five years the other day I thought that you know Frederick you guys also announced as a five-year anniversary I know all of you guys been kind of in this space for five to ten years it's really an honor to have all of you all of you here today and speaking of some of the you know bigger topics maybe compare with a yellow gold there's so many different things and new trends

that happen in the space right talk about you know the rise of large language models probably you know providing whole new possibilities in terms of how we are looking at data all of a sudden there's a so-called proliferation of layer 2 and even some of the applications are moving to the app chains as well potentially this could lead to even more fragmentation in terms of you know blockchain data creating I guess more complexity in terms of you know how you query them um you know I can go on and on and on you know essentially exchange risk and shift to D5 uh you know on search and Regulatory environment all that kind of stuff just curious on your mind what are some of the major conversations you're

having either internally or maybe with your clients um that are that are most noticeable and some of the things that you are focusing on right now then we start with Frederick sure um so we're among other things uh of course as many others leading into llms and AI things um I think it's obviously a huge hype and I think a lot of people assume that you can kind of just like slap some AI on top and then like everything will be great and it will work you know out of the box and be epic I don't think that's quite how it's sort of looking on on the ground and then with your hands dirty um there's a few sort of complications

one is that you need to write like a query a SQL query to interact with the data but you also need to understand you know what's happening on the blockchain what are these smart contracts what are these events and and that's quite complicated in and of itself and I think like a lot of what we do as companies and people do in the community are to like really you know spend a lot of time understanding what what is this actual event like what does a trade mean you know in various systems um and so it's kind of like a combination of these things that need to come together to um you know make llm sort of experience work and that's not trivial yeah um and I think also there's an interesting thing where typically you either have like a very technical audience or you have like a

broader more sort of consumer audience that's not technical yes um and with some of the llm stuff you kind of end up a little bit in the Middle where it's like you need to understand kind of what's the question I should ask right you need so you need some level of like understanding um and I think that's a bit of a unexplored territory in general for like a lot of tech products blockchain data being one of them but basically what is this Persona that sort of knows enough to to ask the right questions um but it's not sort of a pure consumer that only want to spend like five seconds on like understanding something but it's not sort of a cold Rider um so I think that's that's sort of a big question in my mind it's like how many people exist in that category and

sort of to what extent do they still need to understand the fair amount to get value out of features like this or to what extent do you manage to make it like proper consumer yeah yeah so I definitely want to dig deeper on that one because a lot of a lot of nuances and a lot of stuff that I I'm pretty sure that everyone cares is probably going through to some extent as well I mean obviously we're building you know AI power search and we're going through that conversation a lot and when people compare the experience of chat gbt versus some of the stuff that we're going through right now right so it's not so much directly comparable and it's curious like Matthew for example you guys launched the the How's being kind of the conversation so

far with some of the clients uh house being some of the use cases that what's been your observation sure so I'll just follow up from what Frederick mentioned right uh so uh so two pain points right so the first pain point is uh understanding spark contracts so what what we've done is we've built a tool called code reader which basically allows you to take a verified smart contract source code ask it basic questions uh so uh what we see is that still very early use cases folks are basically asking very basic questions like what does this

smart contract to or what are the some of the core functions that you can do and moving on to a little bit slightly different unique use cases we can also see in folks uh modifying a smart contract and then saying what does this do right now after we've modified it so I think that's kind of interesting and going back to the second topic right is like uh it's a basically goes down to from engineering so like knowing what is the right questions to ask and we've actually seen this you know in early 2000 late 1990s research engines right like we're also used to Google right where you just type in something and you'll kind of pretty much guess what you want to ask so I think we're at that stage right now where you actually have

to ask the right questions through proper prom Engineering in order to get the right answers yeah that's uh that's a good perspective and just curious that Alexa is just I I guess this question is for you just from um From nansen's perspective right you guys are obviously pioneering you know own chain analytics intelligence you've tagged millions of you know wallets and people don't I think people tend to under appreciate the complexity of these tasks how you know feasible is it to kind of Leverage AI in some of these and what are some of the challenges that you're going through yeah that's a good question um I think it's true that this is a very

complex problem to solve and there is no Silver Bullet on how you solve it so our Approach at Nelson has always been to combine man and machine right so we have a great team internally of research analysts or attribution Specialists as we call them who are really good at understanding information you might find online for example a fundraising Round And then trying to extract how can I see traces of this on chain and then update our database with the information that you can infer from that information now that only takes you so far it's not scalable to do this uh to get you know 250 million plus different wallet labels

that we have today you have to involve machines as well right so most of the work we've done so far has been heuristics and algorithmic uh based so that means we haven't been using AI as much as we can but we have been exploring things like trying to predict this this wallet likely belong to an exchange based on its transactional patterns yep based on the graph around it and things like that and I think there's a lot of potential for us to do more on that front and my view is that I think to some extent AI might take some of the role that crowdsourcing has done where AI actually

almost becomes a way to tap into the collective intelligence of the internet and so if to to Matthew's point if you're able to ask the right questions and if the machine has the context for example there's a lot of information about wallets uh on GitHub sometimes on Twitter on different Dao forums and so if you're able to fine-tune uh your AI and bring out that context from the web yeah then that will help you in your labeling but I don't think you're going to be able to take the the human out of the loop completely yes I think you need to think of this as man and machine working together yeah I think that I I think it's a great perspective right if I if I come in from some some of my

observations right some of the pattern recognition stuff to your point some of the you know crowdsourcing stuff could be potentially replaced by some of the AI Technologies right pattern recognition identify certain what is potential association with certain things then you have like human element of providing human insight into what's actually happening and maybe you know directly having come session with the clients and people who have Insight that's that's a very interesting Dimension and just referral just anything that you're doing kind of at Glasgow level in terms of how you think about llm and you know some of the latest yeah so so I think my classroom we're using AI um in in two different ways which you know relates very much to what

Alex says uh one is on on on labels and and on clustering wallets right I prefer machine to to talk about it as is machine learning right supervised machine learning or unsupervised one um in that context because you really want to identify um a you know wallets that belong together that are you know basically controlled by the same entity and this is where you get rid of a lot of the noise and and and are really uh you know able to extract um actual economic throughput right if I transfer money you know from my left to my right pocket economically it's not very meaningful right so so you really want to put these things into place um you want to

um definitely write use it for for labeling I think when we talk about exchanges when we talk about these large entities that are obviously very important in this crypto economy to track right uh you you need a very diverse and a very holistic label set right and then the second thing that we're doing is um we're building our own um you know uh query platform as well with an AI assistant um and so um the I believe right the the AI part over time becomes more and more a commodity I'm here you know to Second what Fred said before I think the the real work comes uh with the abstraction layers that you need to build because

you have unstructured data across many different protocols right and so really being able to to make you know a uniform data set that you know an AI can you know basically translate into a query and and surface value is is where you know from our perspective most of the work comes in yeah that's very interesting I mean a lot of people said like bludgeon is all transparent and it's open but at the same time and to your point right there's a ton of pre-processing layers and intelligence that actually go going through it to really understand what's going on and that's exactly you know the values that you guys are providing so kudos to that and just Michael curious I think yesterday or the two days ago as I see like a headline it's in the SEC is now

like leveraging AIS for from somebody's like investigation and things like that just curious what my Genesis perspective what's your thought and what are some of the you know internal conversations or maybe conversation with the clients are having yesterday I think I think one of the things that's interesting when talking a Limbs and A9 General is that AI has been around for a long time and we've been using AI since the Inception of the company and it's like a core part of like identifying transaction patterns being able to identify categories of wallet similarities both from the the graph around them from the patterns they transacting the time of day there's a lot of things that we can feed into to

AI models have been done that for many years I think we all got like a new tool in the toolbox in the AI toolbox that's that's llms over the last year or so and it's been super exciting and I think that there are many ways to use that and I think one of the things that's been most exciting for me is that we have through the customer Network that we built of thousands of customers in the crypto space that use our platform to for transaction monitoring we have millions of data points that's actually ground through data and one of the things that's really scarce and hard to obtain and we all know that like machine learning is all about about having the right training data set and if you don't have a training data set you only get a

get like basically noise out in the other end and we have that very very good Training data set and we'll be using that in other machine learning algorithms in the past and like the very good results and it's interesting right now for us to experiment with that that's one side the other side that we're looking at is similar to what you mentioned before like basically enabling and training the the llm to understand what are the capabilities we have we have what are the apis you can reach into and how can you then in human language describe an investigation augment that investigation understand what's actually going on and I think that will basically enable investigators

using our products to I would say be more advanced compared to where they actually are today and be able to do things that are at a higher scale so very similar to copilot where you have an engineer that would spend hours in in in in trying to look at stack Overflow and find the right solution and now it just basically continues from their fingertips and this is the same things that we apply into our products as well so super exciting got it interesting I mean the other day I should try it I asked us charity VT in your training data set you have blockchain data and they refuse to answer me that question as honestly curious but that is that's interesting so maybe just moving on and you know

just want to touch on kind of Layer Two the app chains and it's obviously very relevant to most of the folks here so maybe starting with Matthew just in terms of you know in this kind of you know some people call it proliferation of layer 2 as all of a sudden it's a lot more kind of infrastructure layers coming out what sort of how do you position yourself in that and what are some of the conversations that you're happy having so um one of the things that we kind of observed and seen demand form uh is L2 chains so everyone's building an L2 chain or an L3 app chain and we've kind of seen this similar Trend a couple

years back so a couple years back it was everyone's building their own evm chain so right now everyone's building their own L2 Now LG chains um so we have an agnostic approach right so if you're building an L2 chain and you want to build an explorer find this so how we build that um and and Alex you guys are probably kind of also pioneering a lot of the conversations and you you guys work very closely with uh you know uh you know a lot of companies who are launching kind of out tools what are some of the conversations they're having yeah I think uh it's it's interesting because the the work that we do collectively here on stage uh becomes uh on the one

hand more challenging the more chains and the more layers you have but also the more because more valuable as well because it used to be the case that you could kind of go on ethereum and that's that's enough uh now we're in a world where basically you have to look at what's happening across so many different chains uh and so I think it's important that we think about how you can create a user experience where you feel like you have an overview of what's happening in the entire blockchain ecosystem and this morning we announced nonsense II uh and one of the things that we're focused on is to create that multi-chain user experience especially in the evm and layer 2 ecosystem because the the good thing about layer 2s and EVMS is

that they happen to have pretty simple uh similar data structures and that means you can also display them in a similar way and so when you bridge you make Bridge transactions you put tokens from the ethereum based layer to arbitrim or polygon and so on you expect I think as a user to be able to follow those transactions easily and so for us I think we're when we go from nonsense one to nonsense two we're thinking a lot more about how you can create a more coherent and multi-chain user experience and so that's something I think users will just expect they don't want to see siled uh have a siled user experience more and more they'll expect to see a

multi-chain view and it goes back to the labeling too because a lot of the labeling work you know that we do it carries over yeah right because you have you know you're using the same address on base or polygon as you're using ethereum and so you can then understand and get a bit of network effects in terms of your labeling your attribution Etc so I think it's a super interesting topic and more and I expect us to we already support you know nine depending on how you measure it 98 of you know blockchain activity in a broad sense you could think of stable coin transaction volume token transaction volume tvl and whatnot yep um but it's super important for us also to not necessarily get

distracted by every chain that launches you have to be mindful of which chains that you actually on board yeah yeah I I Matthew and I were chatting about this uh earlier today basically I think in to some extent like uh we're all uh humble Builders and happy to contribute to the space but like doing a lot of this data stuff is a fair amount of work and I think maybe sometimes a little bit underappreciated like how much goes on under the hood to make all this run smoothly and have all the data quality and performance and all that and it's kind of funky with with all these sort of Trends you see in the space where you know everybody will be at like all of our doors and be like oh you know please integrate you just do XYZ and and you're

good like but but that is kind of like a one-way door decision for us mostly right where it's like if we support something it's really hard to take it away but you know this thing might be super hot today but in in two months no one cares but you have to maintain the thing right and which are you talking about I'm not throwing in one the bus but but I think it's fair to say that this whole Space has you know its uh Hypes and and and and current uh Talk of the Town right so I think that's one of the challenging Parts is like you know supporting all the things that's going to be important but also not hopping on to everything because suddenly you have like a lot of infra costs a lot of operation overhead to support all of it and and that's a bit of a difficult

thing to to sort of get right yeah to chime in here briefly I always like to tell a story of you know five years ago when we started um we actually started to integrate EOS right the ethereum killer no offense in case you know a thing but um but you know these are these like highly unsustainable networks right that that just don't survive right I think if we would have been having this conversation 18 months ago we were probably talking about Luna and and in or you know like other networks that that simply die so it's it's a it's a fine choice to have to do right in terms of you know how breath you know breath versus depth right what do you choose in order to integrate and whatnot yeah I mean I I

think you guys raised that if I win shoes you know I probably constantly fighting against a fight in between sort of uh integrating versus a new chain versus a lot of stuff that you are constantly trying to approach as it definitely kind of echoed that one so just just to chip in on that one because it's kind of funny to hear conversations around looking at like what blockchain to support what not and so on and and you kind of sound my like my product engineering team three years ago and it's like they were all like yeah we need to prioritize this we have EOS coming in oh my God it took us half a year it was painful no way be able to support Solana like have you seen how big it is it's like impossible and so on and we've gone through all of these this

journey and then we basically flip roughly a year ago when we said like this is not the conversation you have the conversation is really there is nothing but everything it has to be everything and the mindset has to be that how do you make a platform where it didn't matter whether you supported something that didn't matter and how do you make an interaction with with the different projects where it didn't matter that it didn't matter and we've been like doing that and we'll be able to to latest launches we did on some of the change we're launching right now are using that platform where it's just gone from like where we would spend months unlike going through blockchains to basically days in terms of like being able to integrate that and look at the data there and

we're trying to bring that down even further because I do think people we are looking at at like as you say layer 2's we are looking at multiple approvaluation of blockchains I actually firmly believe that this is kind of an annoying thing if you if you sit here 10 years from now as well we will say it was wasted effort but it was how to get there I think the number of blockchains will go down long term but it's only going to go down to this awful situation where now basically like a a huge number of blockchains and then only at the end of the tunnel we are back to a few where everyone can concentrate of those but I think right now the only answer is like literally everything which is so painful yeah I mean yeah it's becoming consensus

well at least from some of the folks that I've been talking to you know just that is uh probably to some exact infrastructure Surplus is curious uh they're interesting that you you pointed out maybe just my code that you talked about you know some of the conversations the prioritization I know we live in the world today where probably I'll say regulatory uncertainty is is more than ever what are what are some of the things that you guys are focused on some of the you know discussions that we've been having and also you know from a different perspective or different dimension you know a year ago FTX is still there right and a lot has happened in the space you know Central exchange risk all that kind of

stuff um you know together with regulatory uncertainty and heightened regulatory pressure what are some of the conversations we've been having yeah so I think if TX did a lot of damage and it it was actually in terms of the impact on crypto prices and everything actually not if you look at like we have a Mac Pro winner and a crypto summer in the middle of the window because right now if you look at like where the crypto prices are compared to the NASDAQ to compared to s p it's amazing there's no no one can cry in crypto today because everyone else in the world have like a way harder time than we do so I think that's kind of one takeaway but what did happen is that there was like a lot of institutional players that got scared

around what they saw and got embarrassed because of the engagement they had with FTX and that's like hurt the industry it still does so it would not deny that I do think there's another thing that's very important that's some of the conversation we have all the time and the way I usually like coined that is to say crypto is not about crypto anymore crypto went from being very focused on crypto itself to become infrastructure to augment the rest of Finance to get better rails and that has really happened today so one of the data points that I obsess about is that like 75 of all value moving on blockchains is National occurrences in forms of stable

coins that was not the case a year ago it was not a case a two years ago that's changed so crypto is not crypto anymore crypto is normal finance and it's like growing and growing and growing so this number will be 90 at some point it will be 95 it'll be 99 and then the underlying cryptocurrencies will be seen as like the gas that runs the world to finance which is amazing and then what's on what's what's like on top of that is everything else tokenization of real world assets we just did one that's the simplest one that's cash and that's where we are today and that's some of the conversations we have and see that that's kind of a kind of a very positive thing in in the terms of all crypto where I feel that that's really working

today yeah I I remember either with you or Raphael the other day was mentioning and you know the right metric to look at is not market cap but it's actually transaction volume yeah yeah that's that's that's very interesting perspective and just you know curious Rafael from your perspective you guys obviously tracking a lot of different metrics very insightful what are some of the metrics that you guys are you know focused focused on recently and maybe some of the you know more recent ones that you are starting to pay attention to yeah so um we usually never you know it's usually never one or two right but what we always try to highlight is it's a Confluence of many things right if you look on chain today um from a market intelligence

perspective at least um right you really see this um post bear Market a hangover phase right where there's a lot of apathy and and and and and and and it really um right low volumes on chain um across the board right you see network activities going down you see things like you know liveliness um that is you know a measure for for active Supply is at pre-2021 um levels um so uh looking at different things you see actually that Confluence of um of of your this very common and usual state in which we are which is a

hodler dominated Market currently right which tells you there is high conviction on the one hand but which also tell you everything else got wiped out right and the hodlers right long-term long-term supplies at all-time Highs are the only ones left right so it's low volatility it's uh you know low volumes you know even in off chain you see this right realize volatility as well all-time lows uh Futures volumes as is is low right and I think what needs to happen is um really um you know we need demands in order for this regime to to change otherwise it's it's going to be the diamond hands that that will that will remain until

something changes um yes absolutely very very dry very very dry summer so far yeah but it's it's no right um and and you know it's it's just waiting for for the next wave yeah I agree I mean Alex it just then you know curious from my own chain analytics perspective whatever you know the other day I think you're mentioning real world assets it seems to be booming what are some of the you know interesting stuff that you'll be observing I'm just from um you know own chair activity perspective yeah I think Michael's point on stable coins is really good and really important uh and I think the continuation of that if you think about custodial stable coins is basically things like tokenized t-bills uh which

is going to become I think very big and in fact if you think strategically it might actually start eating up some of the market share of the stable coins themselves uh because you you'll be getting yields and you might be off-ramping the actual Fiat that's behind or the tokens that are represented with stable coins into tokenized tables and that you know takes us further into real world assets um I think one interesting way to think kind of dog coins and pepe and Pudgy penguins and all these different things is that they are kind of proofs of concepts for infrastructure that will really power all of Finance in the

future so if you're building nft infrastructure that works with profile pics and so on that will also work with nfts that are actually real world assets right so if you can mortgage your pudgy penguin well why couldn't you just mortgage your house if that house is an nft in the future and I think sometimes people who are not that crypto native they don't see that perspective and they think you know all this stuff is just crazy nonsense and and cartoon animals um so so I think that that aspect is very important with real world assets I do want to mention in terms of our own Focus internally especially on the product side I think one realization we've had when we think about how we can

help our users better which primarily is crypto investors is that we need to make nonsense more personalized as a product and so if you're a crypto investor you shouldn't be seeing all the same stuff that everyone else sees because that's not Alpha then you get alpha decay and people have different investment strategies and so this is a big Focus area for us to announcement too that you can connect your portfolio and you can track lots of different chains and your Holdings across chains and that allows us to surface signals that are personalized to you and so for us from a product perspective this is really exciting you should be able to track your own wallet segments not just the ones that we create

and so this is a big conversation internally on the product side and I think we should learn from web 2 actually where a personalization has been super important for companies Amazon Google and so on and as the kind of long tail and crypto expands with more and more tokens nfts you need better Discovery tools and I think with nonsense 2 you get a chance to actually personalize what you're seeing in terms of onshine information got it I I like it a lot and especially on the personalization point I think that's that's a great point I also like it that you you don't you don't miss any opportunities to show project hang green you know

very much appreciate that so just um I guess to Matthew just very interesting uh I guess we're just two minutes of time has run out but just curious I know everyone here has been building for a long time and on the last cycle um people are kind of earlier at more or less at the beginning stage to some some of your folks and you know people are finding it difficult to imagine the days when we didn't have that all of you guys products I'm obviously a big fan of all of you guys products but just looking forward what are some of the most exciting things that you are looking forward to and what do you think is the end goal of you know

web3 landscape and just keep it we have to keep it short but just uh you know a few sentences in terms of your prospect and our Outlook that is a difficult question answer so I I guess when I first started out in the space right I've always looked at crypto as programmable money something that you can program and I my my thesis going forward still the same uh but I think there's even a stronger case point for that given the rise of AI so I think AI crypto definitely goes hand in hand uh where we can move closer towards that

future of programmable money and maybe it's Michael yeah so so I think that what I look most forward to are the surprises to be honest because when I got into crypto there are many things I did not were not able to predict I were not able to see that like we will have liquidity pools as we got I couldn't predict like a network project like helium where you could create proof of location where you could create like a worldwide network of low bandwidth and like stuff like that and there'll be some surprises and nuggets on this journey that's going to be amazing so I think the surprises are the things I'm looking forward to uh Rafael uh well we believe the entire world is moving on

chain right and that's what excites us right real world assets you mentioned stable coins it's so sort of you know providing this infrastructure for the for for the for the entire economy right creating more transparency and that's um you know that's the bet that we have at last note Frederick uh I think it's going to be good to have to have some tough times for a period I think forest fires are are pretty healthy and I think the whole industry may be sort of deserved a little bit of punishment for some of the stuff that happened in 2021 so I I think it's gonna be actually super healthy it's going to be brutal and it is brutal uh sort of times like this and it's probably going to be for a while but I think at the end of the day you know the ones that actually really stick to it

and if you just persist you you can do really great things and so I'm excited for for some tough times and then uh we're gonna sort of wonderful deserve the next cycle yeah Sandy with Alex yeah I think long term every asset will be a token uh which means it's going to be on chain and that's very exciting in the short term I'm quite excited about friend Tech I should all try it out you should buy my keys you can sell them too it's okay cool I think we're out of time and it is is I just want to say this again it's great honor to have all the panels physically here today um thank you all for your amazing contributions and insights that's given a round of applause thank you