Surfacing Signal from Noise: Market Insights from Data Experts - TOKEN2049 Dubai 2024

my name is Auggie I'm an investor with CMT digital CMT is a trade uh 28-year-old trading and investment firm based out of the us we've been investing in the space for about eight years we've raised three funds and we've partnered with over a 100 companies in the space including ethery Athena a Infinity coinbase consensus and many others and it is my pleasure to be moderating this panel alongside our esteemed panelist Miguel from Arkham key from crypto Quant and Alan for Doom thank you guys so much for joining us I guess to start I think it would be quite helpful if you could each take two minutes and maybe share a little bit more about your backgrounds and your respective companies Miguel you want to kick us off sure yeah so uh my name is Miguel Morell I'm the founder and CEO of Arkham intelligence uh Arkham is a blockchain analytics platform focused on labeling as many different wallets as we possibly can for the cryptocurrency ecosystem uh the overwhelming majority of our customers are uh retail Traders followed by institutional Traders uh who are looking to do one of two things uh number one attempt to find Alpha in the market uh by discovering new wallets and uh patterns of wallets that are owned by the same type of entity uh and then analyzing their behavior in order to then you know take that in Aggregate and try to make a collective approach of uh what the supply and demand uh market dynamics are in in in the current ecosystem and furthermore determine if the price is going to go up or down and that's what they trade on uh and then secondly uh an attempt to manage their risk right so analyzing portfolios um analyzing you know balance sheets of different companies that they're actually um you know cussing with for example uh in order to to try to protect themselves uh from all the sort of risks that exist in the in the crypto ecosystem um hi my name is key um CEO and founder of crypto Quant we're a data provider um we provide analytics platform for institutional investors for qu quantitative trading so C group um work workan and all these other um major quantitative trading firm um are our in our clients and also we have a platform for retail investors um it's for tracking smart money and monitoring third party risk um which is mostly um by labeling just like arcom but we what's difference from um other data providers is that we aggregate all these onchain wallets um into one single um very intuitive Matrix for tra Traders so we're building Traders community uh who wants to derive trading insights out of the onchain data cool yeah uh hi everyone I'm Alan um I'm at dun analytics we're a community data platform and data analytics platform and um it's basically a place where anyone can build uh queries and dashboards comparing different blockchains different protocols um and then build on top of this vrr API um and yeah we we work with most of the biggest uh blockchains and we provide data for um crypto companies hedge funds um protocols you name it fantastic no thank you and you know I'm super excited for this panel I think you guys should be too and really the reason for that is if you think about it blockchains fundamentally are a database and data structure Innovation and so the topics that we'll delve in today some of the questions that we'll ask will be where we're seeing the most Innovation at that data structure uh layer whether it's things like data privacy the impact of AI on data data ownership uh data ingestion so with that I think let's get into it maybe starting with uh data privacy right there is at least today a fundamental trade-off between uncovering or surfacing data and maintaining data privacy you know I think on the trading side it's it's quite apparent right where if I'm doing some activity on chain you as the third party Observer would see exactly what trades or pools that I participated in which isn't necessarily ideal right uh and so there are areas looking to address this space uh categories like fhe where you can uh compute over encrypted data mixers obviously where you can obfuscate flow but I guess the question for our panelists really is you know this tradeoff that I've been describing is this a given in blockchain something that we need to accept or do you see a world where both of these goals are achievable together I I I can go first um we we tend to get this question a lot um so you know the the interesting thing is I I I think that fundamentally uh blockchains do in fact have this uh this Paradox where they are both uh you know simultaneous ly uh sort of meant to be very transparent uh immutable and and very auditable um but at the same time they also have the sort of uh inherent characteristic of pseudonymity right and so when when you build a a piece of technology that is both uh meant to be incredibly transparent and auditable um but at the same time by default um you know gives the user a sense of kind of a sense of privacy via pseudonymity um um I think you basically then end up attracting the kinds of people and the kinds of users who like both you end up getting a a huge Camp of users and of people who are like I love crypto because it's super transparent and super auditable we should be using blockchains in order to make sure that you know this data is legitimate that people cannot mess with it uh and in order to provide a certain amount of data security um by the fact that you know it's so so difficult to for example go back and you know alter the Bitcoin blockchain for example that data is there it's there permanently you know once the the block has been mined you know it's nearly impossible to go back and and edit uh without reaching consensus and so you know but on the other side you then then end up getting all of the sort of privacy type folks um who who want to use the the kinds of use cases and applications that um you know are yielded from having a kind of Technology where somebody can transact in a financial matter um you know with with pseudonymity right and so I think the answer to this is that both of them exist in parallel um it's a little bit paradoxical um but to answer your question I do think that they will continue to exist uh in a very similar kind of way uh for the foreseeable future um and so I I my vision for the way you know cryptocurrency is going to go um is that it'll be very similar to kind of like the broader internet right the broader internet um is very um you know transparent people you know use their real name accounts on Instagram or Facebook or Twitter or you know whatever um those who choose to be pseudonymous can can do that right there are plenty of sort of anon accounts and Anonymous kinds of tools like tour and VPN and other sorts of services that people can kind of opt into um and I think that cryptocurrency will will go the same way I don't think that privacy tools um or or ways of people maintaining their privacy will disappear I think people will continue building that kind of technology I think those who are interested in using it will continue being able to use it it's not going to go away permanently um and I think that that's also how it should be if you if you want to opt for a more kind of Private Financial setup you should be able to do that right um and and and people do do that within the cryptocurrency industry um however I I do think that that will continue to exist in parallel with the other fundamental characteristic of blockchains which is transparency and auditability I think the market will decide the technology adoption so there's a Monero you know zcash all these ZK um fully um you know cares privacy for your transactions but you know some people use it some people don't so I think um market demand will tell us what to build for example if cbtc and stable coins it gets adopted you purchase orange juice in the convenience store right and and you can basically aggregate demands for the orange juice around the world in in every 15 seconds by by block time right so in if you if you think about it um once we need to calculate um orang juice supply and demand to calculate interest rates for the world or you know the nationalwide Matrix people will put some data there and people will exclude some privacy data um by it and it all comes from the demand I think um yeah that that's it um yeah I agree like it's I think it's nice that there's different Alternatives but I agree with Miguel's point about like there's a lot of transparency and um auditability of this so I think what's cool about this is that you can actually um um you can see what data you have left traces of on the blockchain and then you can kind of choose if this is something you want to continue doing or not so you can choose to either keep participating in that blockchain and leave that data behind or you can you cannot do it or move to something that's more kind of privacy focused with all the Trad offs that that entails uh and I think this is a big shift from like my data on Facebook or Instagram or Twitter or whatever I like there's gdpr are but I not like I don't really know what data they have on me and how they're using it um so I think that's that's kind of the the most interesting part about this that that um I could see potentially what they have on me and then I can either not participate or to participate yeah I agree with key I think to to add to that point it definitely is true that I I sort of I I like calling it you know the Market's revealed preference right there are plenty of people who make these claims online about you know the people want X or the people want y uh and really what they're trying to say is I think this is a good idea and I think this is what the industry should look like it actually has nothing to do with what everybody else thinks um and I think that that that really shines through in these sorts of things like you know if people want to use privacy coins privacy protocols you know privacy tools Etc they will go and do that and that will show in the market the reality is that you know things that are you know Ultra transparent uh are just way more successful that's just the the the reality right if you look at the tools you look at the coins Etc all all of the things that are you know the largest things in the market you know most of them have absolutely nothing to do with privacy right and that's just the Market's preference right um and and that's not even talking my own book that's not talking you know our preference or anything I I I think there are other things that for example I don't agree with um that you know have become Ultra successful like the people want meme coins it's just like the truth like these things have insane Market craps they grow like crazy and the reality is that people care more about investing in like cat memes than they care about privacy because those things have way more of a of a higher market cap and more trading volume right and so I think the market for Better or For Worse uh does in fact show its preferences and and what it thinks is good ideas or useful or important or whatever um compared to other sorts of things and uh you know the the transparency Camp right now is winning makes sense makes sense okay I want to maybe switch gears a little bit and unpack one of the comments that Alan made because I think it dovetails very nicely into another topic that we should discuss which is data ownership right uh you know I think there are companies that we've seen in the space working towards giving individuals ownership of their data and the economics associated with it right even within our own portfolio we've invested in cookie 3 and Snicker doodle which are at a high level working towards this objective right and you know in my opinion it's a worthy objective but sometimes feels a little bit far out and so the question I have for you all is where do you think we are in the process of providing individuals with ownership of their data and frankly how realistic is it given the backdrop of web 2 business models today um tough questions but I I specifically remember um when arom launched um people some people talking about their platform um product is really great but um people worried about toing people um actually I my address is also doxed in on arom as well at first I I feel like oh my privacy got bridged you know I feel I felt bad but at the same time you know just like market demand uh okay I I'm going to split my addresses for privacy but also for the you know disclose to the public and Aram did really good job uh in terms of the transparency so when someone you know some insider trading or um some proof of reserve and these These are really important and advanced industry and we once we do like uh these type of experiments in onchain data um Market will decide what to do when don't and eventually people we we basically educate the importance of the Privacy for each individual um people for now I think most of people they don't understand why privacy is important why bitcoiners and other crypto people care privacy but once they Fe um you know feel just like my in my case um they will car the pr privacy and onchain data provider and platforms we our role is very important in that in the perspective um I I think one interesting thing here is like so my individual data might not be very interesting to company but if we're like a thousand 10,000 people uh organized somehow and and we together kind of want to sell our data and we opt into that I think that can be an interesting way of doing this um and I don't know maybe it could be organized via some Dow structure or or something like that um where you have this kind of almost like a data Union or something where you where you um you explicitly sell your data to someone but you know who and uh you get compensated for it as well uh so kind of flipping the web two world where you were they are just selling your data and profiting from it but you don't really benefit from it uh so yeah I think that that can be an interesting thing to explore in the future yeah I think from from our perspective um we've taken a really interesting approach which I think has you know caught a number of people off guard which is that in building our platform and in releasing all of this data we've at all points in time tried to make the data as accessible as possible right in the beginning of of of my business you know I I had to make a decision okay are we going to be kind of like a uh you know a SAS kind of company are we going to be like a B2B uh institutional kind of company um or are we going to do a kind of you know uh direct uh consumer application uh where we try to get as many users as possible right um and and My Philosophy as an entrepreneur has always been that you know the value of your of your company roughly obviously there's speculation and other sort of market dynamics at play but but on a very high level the value of your company uh is directly in proportion to the amount of value that you're producing for the world and for your user base and for your customers right if you look at all of the largest companies in the world they have a dramatically outsized impact uh on number of people uh in terms of value that they're producing for people uh and and what they're bringing to the table so uh I think in terms of Arkham we we've been pushing that as much as we possibly can and trying to deliver value to the cryptocurrency ecosystem uh by building you know some of the best tools by providing some of the best data and by making it as accessible as possible for everyone right um many other companies don't take that approach right there are plenty of companies that uh do Data Mining and data analysis and then they're just sold to you know a small kind of cabal of Institutions or government Regulators or law enforcement or something like this um and I think that there are now more blockchain analytics companies trying to break out of that mold uh and not just trying to you know sell to the government or something but rather actually providing this transparency and auditability to the community in order for them to try to keep themselves safe um and I think that that's been sort of our part in you know at least attempting to make for example people aware of the sort of information and data that already exists about them on the Internet that is most likely already being used by other kinds of Institutions for more private purposes uh now in a kind of public Manner and you can take further steps in order to for example if you're interested in offis skating or protecting your future wallets now you know better ways of of of being able to do that so that that's my kind of General read in terms of whether or not people will be able to own their data in the future I think it's going to be very difficult as you know if you're using somebody a platform that a company has built uh that they're trying to extract value out of um and they think that they can sell it in private or something like this it's it's very very difficult to try to retain you know ownership or financial royalties or something uh of your own kind of uh of of information because it goes against their incentives um we have no reason to to kind of sell this sort of data in public or in private because it's all public anyway right like anybody can just go on Arkham and like everything that you see there that's what every single one of the users gets that's what we see internally um and everybody's getting it for free right there's no there's no back Channel there there's no point right um and so I think the more companies can kind of move in that direction I think the the better and healthier uh the data sort of ecosystem is going to be makes sense great um let's move on to another topic topic guys which is you know it would not be a 2024 panel without bringing up AI in some form or fashion uh and you know the maybe the the broad Strokes of the question is we're obviously in a position today where AI is making Leaps and Bounds forward at a very rapid Pace uh you know we ourselves have seen many companies that are attempting to productize AI across the value chain of blockchain we invested for example in a company called hyper native they do it's basically an advanced AI monitoring solution for detecting vulnerabilities in smart contract code but there are many other use cases and so I'm curious for you guys how are you currently utilizing AI in your businesses and how do you see it impacting your longer term product road map if at all um for now uh we've been using AI for to um identifying addresses so if you build a model that determines whether this is exchange or not um it's its accuracy is around 85% um is so um that's the that's one thing and secondly I think AI really helps us to um you know interpret smart contracts uh so we whenever we build build the abstraction tables for um certain smart contracts we ask AI to decode all these data um and lastly I guess unchain data um is is for transparency But Ordinary People they don't understand how utilize launching data and how interpret those data um to get get the knowledge knowledge they want to um they are curious about so we're building anal analyis community and based on this um onchain analyst Community their contents um I I guess we can train some kind of analysts AI analysts um but it's accuracy is not like really helpful for now so I guess it it will take 3 to 5 years to make it actually helpful for now it's like um just uh for checking whether this is fact or not um but it's hard to make those AI models to become more financial advisor so we um recently released a product um that's called Dune AI it's basically so our main interface uh has always been um query like SQL writing SQL queries and getting that data that way so with with dun you basically chat with um with our data so you can you can just type like what what was the volume of uh Dex trading in the last 24 hours or week um so I think that's that's kind of a cool First Step we've taken in making our like using AI to make the ux better um and our mission is to make crypto data accessible and I think that's that's kind of a like I don't think everyone will be able to write SQL queries but almost everyone will be able to write sentences to to get data um and then I think more long term um I think AI can help us with discoverability as well we have over a million tables that you could query so just like doing normal search can sometimes be a bit challenging um and then I think another thing longterm um Rec recently partnered with snowflake to put our data make our data available on snowflake uh so I think other people using blockchain data to build models I think it's also still um still kind of early and untapped yeah yeah the majority of our of our usage of uh you know anything kind of AI data science machine learning related uh has been uh with the labeling um you know with with our platform for example um you know we have millions of entities and you know over a billion labels that that we need to manage you know that that cannot be done manually right you need some uh you need some automated assistance to be able to do that um and so actually ever since the beginning of our of our company um we have been using you know different sorts of uh machine learning tools and um and and and AI in order to do uh for example automatic uh categorization of wallets you know on the most kind of basic level um you know certain wallets fit particular patterns right like deposit wallets will always uh later transfer to uh you know they'll receive money from like a gas deployer and then they'll always transfer to a hot wallet um and so these are the sort of automated kinds of things where you can do things such as deposit wallet detection uh and then you know if you're able to detect that something's a deposit wallet um but the hot wallet is unlabeled then maybe you can then use that to make a prediction about what exchange it belongs to right based on the on the pattern of the US users or the wallet surrounding it or the other hot wallets that it's interacted with Etc and so if a human had to go uh and do that across every single wallet on chain it would it would take an unbelievable amount of time right and so um this is one of the use cases uh for us of uh of different sort of machine learning tools and uh and artificial intelligence um we also have a a frontend uh kind of more user facing application for it as well which is called the Arkham Oracle uh where people can do similar things where you know you can type basic queries in about things that you're that you're interested in if you have difficulty searching through our platform for something in particular that you're looking for um which has the ability to you know answer questions and respond like this but I agree with key that we're still a pretty long ways out before you know these scenes can start giving like Financial advice or something like this I think it'll be a while not Financial advice got it yeah cool um let's move forward guys and I'd love to talk a little bit about data and suggestion right I think a lot of you guys obviously do a lot of work when it comes to onchain data um organizing querying labeling things that have come from in effect the core consensus you know I think another interesting problem in data is how do we ingest real world data from offchain to onchain because I think that presents two very unique problems that need to be solved one is ensuring the validity of that data through things like let's say ZK ml and the other is how do we ensure the trustless of our data oracles through for example IG layer abses and so I'm curious to get your guys' view on this topic how do you see the ingestion of real world data evolving and how do we solve these two problems how do we verify data being submitted on chain in a trustless manner that is a really tough question um I think that uh you know this this is something that we often have to solve for because yes there are occasions where you use exclusively onchain available information in order to for example find the label of a wallet um but oftentimes at Arkham we also have to use offchain information in order in order to do that right you know going through uh things like court filings going through you know Twitter or you know other sorts of services and so you know that that is a situation where being able to do that offchain uh data sourcing uh in an automated way where we can verify the validity of the information that's a problem that we still need to solve for right that that is a problem where you still need to like hire people to sift through the data analyze it themselves verify it themselves Etc it's not something that is yet automated or done by the community or done by a decentralized Oracle or something like this though it is in our plans right and it is something that we have you know publicly stated that we are interested in doing because it would obviously be a massive benefit to to our business and to our users right and the ecosystem uh as a whole but I think it it's incredibly difficult um and and my response to that is you know that problem of sort of in a decentralized way being able to bring offchain uh information because of the fact that you need to verify it and and and and validate that the information is correct um this is something that is still needed to be done by a human human uh and cannot yet be done in a decentralized way in a way that's not um for example susceptible to you know fraud or you know attacks uh right like you know some people say well you can put the information up online and then have people like vote on it um but then how do you how do you not get manipulation for example right and so there's a lot of these sort of uh edge cases and other kinds of problems that that mean that it's incredibly difficult to do this in a decentralized manner I think it's uh almost impossible to make trust list um for offchain data uh so we we might need some kind of semi um trustless I mean not trustless but decentralized way so in order to achieve this level I think the future will should be um we need to GA gather uh make blockchain adop get adopted for example you know the way that all the alternative data providers um gather the data for the consumer you know index or other um indexes they just call to uh you know uh for example chicken Factory and then ask ask about hey uh how many x x um hatched today I mean uh get get today and then they just put these numbers into Excel file and they gather the data that's and it's it's lagging and it's not like very scientific right so um in order to get those offchain data for example once stable coins cbtc Bitcoins these on P2P payments are you know pretty widely get adopted they can do some kind of smart contracts and then put the numbers um correctly if they don't put numbers not correctly and it it just like um got founded um we we can just slash them based on the smart contracts for example um so I really believe the future of onchain data will be like change the monetary policy um around the world um and offchain data can be trustless if we get them into a smart contract and then build some some kind of slashing policy yeah I think um I agree this is not a diff not a easy problem it's a difficult problem to solve to like I think in general to be able to trust data um I think in our case like a lot of we rely a lot on our community to um like to build models and to to for example um submit labels and stuff um and I think that's one way of of trying to validate data uh anyone can make a mistake but then anyone else could uh also say that the other person made a mistake so you have this sort of um very rough social consensus on like that this is correct or not uh and I think transparency is is important here was like okay if you have any data where did it come from and why did they publish it at the time that they did and so on uh and I think in that way you you slowly start building trust um and then you can choose if you want to trust a certain provider or not um so yeah I don't think it's I don't think it's um an easy problem but I think there's there's ways to like slowly um get there and I think given that it's permissionless as well like to to sort of publish this data on chain if you don't like what chain link is doing what whatever you can you can just publish your own data somehow and and start building up something that way which I think is um also quite interesting Mak sense great uh we're just coming up on time here guys so maybe one final topic uh that we can discuss is data in the context of this multi-chain world that we live in right and I think the the backdrop here for the audience is you know I I think it's fair to say that historically evm has been the def facto standard when it comes to data but with the rise of Solana and others um data standards may need to change and and frankly for an application layer developer new problems arise where different chains meet Silo data means different data standards means event logs being emitted from these different chains they may not even cvent logs whatever the nomenclature is for the respective chains the point is the data problem data structuring problem becomes that much more complex and so I'm curious to get your guys' view with the rise of Solana and others how do you see data standards changing and I guess for a for for a Dap devel building multi-chain how do they maintain a holistic overview of their data across different chain environments yeah I think think um I think Salon has really shown that you can um sort of there is an alternative to the evm that people really um care about and I think this kind of creates the potential for more of these Alternatives like other like solanas basically that are not that that work in completely different ways and have different data models and obviously this makes our lives as like data that that are doing data injust much harder um but I think one thing that that we can also try to solve is to try to create a kind of unified model for what is an event log or for example what is a trade on a DEX how do we Define that uh cross chains um and I think there's there's many of these cases where I think it's our job as U analytics platforms to try to make sense of all of this and create some sort of standards uh cross chain so that you can actually compare things um between all of the platforms I think it's a great question um whenever we expand our data to other chains or um new altcoins decoding smart contracts it sucks I mean it's too many and there's no standards hey you're tax and you know how we can calculate your your tax revenue right and the the way that Cal to calculate their tax revenue is different from others every tax tax have different protocols so I think some um leading oning data provider maybe dun um should do some kind of um build some kind of onchain data Alliance and then you know say hey this is just like erc20 you know hey this is um standard to calculate tax protocol revenue and then data providers like us um we we can calculate easily more easily um those protocol Revenue because um the the way to calculate these aggregated Matrix differenter from on data providers so some data providers say um you know here their product revenue is three billion but others say it's two billion um this discre discrepancy comes from just standard how to calculate um aggregate the data out of the smart contracts so uh we really need I think it's it's time it's the right timing to do some kind of like building standards for the industry yeah yeah I think for us you know we started with ethereum um and that we initially only supported ethereum in in in the beginning and then over time we have added significantly more uh evm and non uh evm chains as well uh broadly speaking you know what we've tried to do is analyze where the majority of the real onchain activity and volumes are happening uh and then use that as a kind of guiding star for which chains we are going to end up supporting next because as Keem mentioned like it's a huge pain um to actually add support and there's ongoing costs uh and there's all of this maintenance that needs to do and all these different chains end up getting upgraded and Etc but uh the reality is you know when I first started the company uh you know ethereum was uh you know kind of like one of the only shows in town in terms of where a bunch of the activity was was happening obviously besides Bitcoin um and now that has grown and there are other chains that have their own you know massive communities that yell yell at us on Twitter and stuff in order to add their chain um for them to do do analysis right and so you know it's job to try to provide this uh this information to the largest amount of users that that we possibly can right um and so you know from my perspective I do continue that I I do believe that crypto will continue to be multi-chain uh and that more and more of these ecosystems will will pop up um I'm not sure that we'll ever get people to agree on following a standard process um and likely that is a you know that's a stone that us data providers will have to you know keep hauling around but um but I do think that you know the more of these that exist the better the more sort of antifragile our ecosystem becomes uh and it's just up to us to to build as good of tool as we possibly can makes sense well with that guys we're at time thank you so much to each of you for joining us and for those in the audience please join me in a big round of applause for our for our panelist thank you