Data Wants to be Free / David Mihal

EthPrague 2022 · June 2022 · 27:03

David Mihal argues that data wants to be free, starting from the point that a decentralized financial system needs data metrics to be understood. At ETHPrague 2022 he reviews the current state of data in crypto and asks how the ecosystem can make sure that its data is trustworthy and easily available.

Speakers

Topics

From EthPrague 2022, hosted by ETHPrague.

Transcript

foreign what's up everybody how's everyone doing so yeah my name is David um I want to talk about data how it wants to be free how we can disintermediate web3 and the the data around it so um before started I'm gonna say kind of who I am I'm a developer I've worked on a just a bunch of different uh kind of crypto blockchain projects over the past year also helped organize liscon which was a big conference we did in Lisbon back in the fall that's me as a lizard

at the lizard conference um and yeah so this project uh sorry this talk is related to kind of one of the main projects I'm working on now which is crypto stats uh kind of project for open data metrics but before I talk about crypto stats I wanted to kind of like zoom out a bit and talk about you know what's the goal of blockchains in general what are what makes blockchains really valuable and you know I think if you ask most a lot of people here a lot of people in crypto in general let's say decentralization um which it's not wrong but yeah I see decentralization as um it's an attribute of blockchains but

it's not really the goal like decentralization gives us other things and one of the the most powerful things decentralization gives us is uh disintermediation you know the ability to remove uh kind of a central party from from something I think I have a definition yeah removing control from privileged entities um and yeah just to kind of give some examples I want to go through um you know the benefits of decentralization or sorry of disintermediate disintermediation and how it's already applied to um kind of what we've done so far in in crypto um you know when we disintermediate protocols or or things we we get

neutrality you get you know predictability which is really important for you know if you're going to build a business on top of a protocol um you remove counterparty risk you know the having to trust someone with your you know your money or whatever um so yeah let's go through kind of like how crypto's already disintermediate removed trusted intermediaries you know and by the way I picked um the graphics from cointelegraph because they have like the coolest Graphics so here's a sexy Bitcoin getting painted um you know Bitcoin the the original blockchain application um you know it's it's so powerful because it disintermediates kind of the power of the state to to create money

you know this intermediates from technology companies you know we kind of almost had the the potential for like Facebook to make money and and you know Bitcoin removes the need for us to trust them um and you know bitcoin's always compared to Gold but even gold you need to like if you want to move physical gold you need to like hire companies and take them across like Customs you know controls so Bitcoin removes the need to do any disintermediation in terms of value transfers and that was the original really powerful idea of blockchains then we had ethereum come here's ethereum talking about himself on a television show um you know ethereum disinterminated the

the need to kind of trust uh third-party entities to kind of like run code for you to have this Unstoppable code um that you could you could create agreements between two parties and not need like a a trusted third party to um to like arbitrate those those decisions um you know it gives you property rights as I say this is if we think of blockchains as you know kind of Nations these like digital Nations uh property rights are super important for a nation Nations that have strong property rights have strong economies you can invest in a country with uh strong property rights so you know really like trustless decentralized blockchain lets us have

those properties which lets us build these amazing economies um and then we can start building applications on top so you know I think kind of the the most simple but most powerful example of something built on a blockchain would be a a decentralized exchange like such as uniswap so uniswap you know I think we kind of forget how like game changing unit swap was because now dexes are everywhere but you know if you go back to 2017 if you wanted to make a token and you wanted to buy and sell that token from people you had to go to a you had to go to binance you had to go to one of these centralized exchanges and they would charge like you know fifty thousand dollars just to list

their coin on their list your coin on their exchange so with uniswap all of a sudden anybody can sell their token anybody can make a liquid Market to buy and sell um you know the government or company or whatever can stop you from you know buying and selling tokens so now we're kind of disintermediating removing some of the need for the finances and the coinbases of the world um and then yeah for ens I I don't know why there's um I'm blinking on this guy's name Mark Mark Cuban in front of a swear ens name but um you know I I wanted to kind of put an example of you know kind of where we're getting to now on

um how blockchains are disintermediating the world and one way they're doing that is um nfts and I think ens is a great example of an nft that's you know already has more you know a lot of utility a lot of power and has also has a parallel in the existing world you know what is ens ens is a domain name everybody understands the value of domain names but it's a domain name that's not controlled by you know one you know one organization because the DNS system is still kind of controlled by the US government at some level so with ens with kind of nfts like this we're removing the need for parties to like issue and control digital assets

and letting people create like this digital real state um so you know again I can kind of see what we're doing in blockchain is building up this stack of um you know adding layers and disinterminating more parties giving control you know taking control from centralized parties and giving it to either the the community at a large or to decentralized protocols and the thing that is really interesting to me is what comes next what's the you know we've disintermediated a bunch of like sectors already what can we disintermediate now so this brings me over to kind of the the stuff I've been working on for for a

bit which all started with this website called crypto fees I guess it was crypto fee stats you guys know this site this is like the original version it looks a little bit different now it's still pretty similar but I originally made this the ugly version of this back in like 2000 sorry 2020 in like the middle of D5 summer um when like all of a sudden like I noticed that uniswap fees were becoming like competitive with even like Bitcoin I was like wow that's really interesting that's a an underappreciated metric so um I made this site you know I'm still kind of running it it's become popular as one of the the most popular places to look at fee Revenue

um but running the site I kind of like again like I made the site as just like a side project but I was you know running it it was getting popular and I started to kind of understand some challenges for running a site like this so first of all like every oops every D5 protocol was like reaching out um wanted to be listed they wanted me to get their data and so like you know there was a lot of work for me to kind of like work with every D5 protocol to get accurate data for them on our website um there was disagreements about the data you know I've gotten in some Twitter Wars with people sometimes because they go oh you're you know you're you're under counting our data and you're only doing it because you'd want our token to go down or something

like that um and finally like people started reaching out companies big companies small you know hackers whatever they're like hey you have really cool data we want to put it on our website we want to make new stories with it like how do we access this data and so I felt like all this you know power was kind of running through me to run this site and I kind of didn't really want that like I I just wanted this site to be a public good that the community could could look at and get access to good data so so I was becoming a trusted intermediary so the problem I wanted to solve was how do how do I remove myself from that position um you know what are the ways to kind of uh be a data product like this you know I I

didn't want to um there were a few approaches I didn't want to take with the site um here's a screenshot from from nanson which almost I don't want to be like knocking on Nance and it's awesome site but who can pay through thirty thousand dollars for a data website like um this is a kind of it seems to me like a web 2 data model which is we're going to take data which data again you can share data for free that doesn't cost you anything to like give somebody data but build a wall around it and say oh if you want to get past that wall you need to pay a bit of money to do it so I didn't really want to take that approach because I mean the reason crypto fees got so popular in the first place was user contributions being an open source

project um the other way that that can be popular uh to kind of monetize something with data is to um you know kind of make it a part of a bigger organization so this is coin market cap a screenshot from coin market cap which you know was a really still is like one of the most popular crypto data websites and a couple years ago they got acquired by binance and shortly after they got acquired by binance BNB became the top defy token on their defy token list and you know I don't think BNB is uh you know a defy token on the same on the same scale as you know chain link and maker and compound especially not at

that time I don't even know if uh Finance smart chain had even launched at that time so seemed like there may be whether this was intentional or not like you can see that there's a a financial interest in binance using a a data kind of property that they have to promote their own projects so you know kind of being sponsored being controlled by a larger profitable company didn't seem like the right approach um but I think we're seeing like this new approach a new um category of projects that are trying to take like this this other direction on how you can make data available and one of I think the best aligned projects is a site that I'm sure a lot of you guys have seen defy llama

so defy llama is this awesome website it's probably the most popular website today for tracking total value locked which is one of the most popular data metrics and they've taken the approach of saying it's going to be totally open source it's going to be neutral you know this data is a public good this is something that people who are you know in crypto in D5 and web3 like they need access to this data and we want to find a way to to make that available um included this really like good screenshot of Xerox and GMI the kind of the guy behind D5 llama um you know what was he saying here um people going oh like why are you adding cardano to D5 llama and he's like We'll

add anything like it's not it's not a data site to go what project is good and bad that's you know that's the individual's kind of uh decision to interpret data however they wish so so what I'm hoping to do is uh you know I've created this project called crypto stats where I think we want to take this idea of kind of Open Source Community run data and take it to the next level um so what we do with cryptostats this is kind of where I see us in this pipeline of um you know how data goes from being you know data on a blockchain you know data all starts like on a blockchain itself you've got the blockchains themselves you've got D5 applications basically you've got a bunch of transactions

events calls like raw data next step is you need something to index it you need um you know nodes powerful computers databases that are going to like comb through all this data organize it into into some schema so you know the biggest one by far is the graph where we're huge you know fans and users of the graph protocol um there are some things the graph still can't quite do yet we still pull data from things like coinmetrics which is like a more centralized company that provides data we pull some data from xerion things like that but you know this is I think one layer which is the kind of the infrastructure layer of um you know aggregating blockchain data and turning it into something that's

more easily accessible and then is where crypto stats comes in which is you've got all this data in different places you've got different subgraphs you've got different websites API endpoints things like that it's in different formats you need to kind of take that data do a bit of aggregation you know not like batch aggregation not combing through like tens of thousands of transactions but do some aggregation and give like a normalized result so that you have uh you can take any query for for you know any project that's supported and get like a output that looks the same for any project so that's where cryptostats fits in and then what we you know try to do with this is make it available to anyone at

the presentation layer like the the the last layer before data gets the consumer is someone making an awesome website some awesome data visualization some some really cool apps something that presents this data the user and you know tries to figure out the best way to communicate it [Music] um and so our goal is to to make all this data open and free and hopefully have like you know lots of competition of people about who can present this data the best way they can um and so like I'll dive a little bit more into the like the technical side of how we try to do this uh you know what crypto stats is right now it's it's quite simple um every protocol on on crypto fees is

in these you know data collections we have uh and each one is represented by an adapter so this is like some sample code for the the tornado cache adapter and so adapter is just a piece of JavaScript it uh you can see this one here is doing a query to the graph protocol we kind of Define a graph a graphql query we query This Tornado cache sub graph you know give it some dates and block numbers whatever we end up kind of subtracting the days from the the fees from one day against the fees from the other day and then that's like the fees for tornado cash for that day so I think like pretty simple piece of code we've also got a bunch of metadata at the

bottom um so the goal is to have a bunch of these uh adapters uh they're all you know stored on ipfs uh kind of indexed on chain and that way like you know anybody can kind of pull what are basically instructions on how to get the best data and you know I think of it kind of like Wikipedia if somebody sees a better way to a better data source or a better way to calculate data they can come in and and update that um and then here's the kind of the the internal process of how a website would use crypto stat yeah crypto stats data is that um as I said there's kind of like an

on-chain registry of these adapters for each collection so you kind of query those adapters or require the list of adapters which are you know um just a bunch of ipfs identifiers get the adapters from ipfs those adapters kind of execute the code to like reach out to the graph or wherever the data source is and then you kind of bring that data back to your website and you can display it however you'd like um but we have like a JavaScript SDK that does all that for you so here's like 10 lines of JavaScript for um how you could get like the the crypto fees data like the fees on a certain day with again I think like a super simple

amount of code um yeah here's like kind of uh one of these collections so it's a bunch of you know a bunch of protocols bunch of uh data and as I mentioned like you know we want anybody to be able to contribute to this this data um you know as I was saying before I don't want to be the one going in building every adapter for everything that makes me like the central point of failure so if anybody can come in they can read the code they can understand how data is generated and if they want to edit it if they see an improvement that they can be can be made we've got this like really cool editor it's like

an IDE but it's right in your browser so you can go in there update the adapter and test it out and kind of publish it back to the community so that the community can update it if they if they need to so that's kind of where where the Project's at now um but you know the goal is to to not just like provide these adapters but to provide any tooling or infrastructure that's helpful to making data open and available for everybody so these are kind of some other we're working on some things with the graph protocol to make it easier to build subgraphs we want to support like more than just the graph protocol like any indexer that exists

um we want to kind of support uh people who are building websites building data sites making it so that they have the tools resources and support needed to like build really cool websites um and you know eventually like the goal is to kind of turn this into a dow to really kind of give the governing control over like where you know which adapters get chosen and how to fund you know get distributed to data projects uh into a dow itself so if uh if you thought any of this was interesting if uh you kind of are interested in our vision of like a world where you know all the the crypto data uh is is open and free to everybody and you know

everybody is contributing to it uh would love for you to join us we've got you know Discord all the usual kind of community stuff we're looking for uh technical people to help build this out and of course non-technical people as well to help build this uh community and if you're hacking this weekend and you want to You're Building something data related whether it's a site that you know generates data uses data presents data then you know would love to have you building on our stuff so so um yeah if you're interested in that reach out to me come find me around here or you can reach out on our Discord or something like that so uh yeah thank you guys so much um I guess we'll open it up to questions

foreign okay there's a there's a mic for questions from the audience does anyone have any questions raise your hand and stand up and we'll get a microphone to you okay we've got a question in the front row what is the iridow strategy what are you already doing to implement that can you unfold a little bit there in terms of um doing what's how it's happening um and that part is still like not super planned out you know we have a channel in our Discord where we've discussed like the different Dow tooling and stuff like that

um you know I definitely want to make sure that uh I think kind of like structuring Dao's is an area that there's been like not as much kind of innovation as I've hoped so I'm excited about stuff like um element finance and optimism like some of these projects that are trying structures that are more than just oh we're going to do coin voting um you know things that do kind of delegate power to to um different groups to individuals if necessary do it we've talked about stuff like um optimistic governance kind of the idea that um you know because again most people are going to behave correctly the first time that you can kind of delegate people

power and have some recourse in the event that they behave poorly you know obviously this is similar to like how optimistic rollups work but um again element Finance is one project that's trying to apply that to governance so um we've got some cool ideas but again it's still kind of an open discussion for the community to figure out okay we've got a question in the middle hi so is your data source currently mostly depending on the graph does that mean if graph sings graph does the graph protocol doesn't support you that's a pro kind of constraints of what you cannot correct so I think most

you know the golfer kind of having you know open trustless data is to pull from the most open trustless data sources and I think right now like usually that's the graph you know um the graph is like most subgraphs are kind of open anybody can you know kind of run their own subgraph if I want to run like the uniswap subgraph on my server I can do that I can change it I can view the code you know they're moving to this decentralized protocol where it's not even just the graph that's running the servers so um but you know I don't we don't want to be constrained by the graph as you said so you know there's a lot of data for example like data from the cosmos ecosystem I know the graph is adding support but like you know generally like we can't get

data from the graph for the cosmos ecosystem so we can pull from from any you know endpoint any HTTP endpoint so uh if it's centralized decentralized like we can pull from wherever but kind of the the goal is to find the most trust you know trustful but also trustless uh sources of data to pull from okay we've got another question if hi and just one question so if I understood you correctly your adapters currently are only geared towards cryptos fees right so if I have a sub graph which is from my application which has nothing to do with D5 which is some real world application and I I want to pull data can I use your adapters or

do I have to write my own adapter for that so if you you want to pull data that's like unrelated to like crypto or D5 to D5 stuff so you can build I mean anybody can build any adapter like you saw we have this editor you can build an adapter like if you think it's a useful tool uh you can build it and you can publish to ipfs and you can kind of import it and into your site um so there's adapters which anybody can create and then there's like this adapter registry which is kind of the r curated list of adapters that like we as a community think are you know trusted and accurate so there's no reason you have to use our adapters you can make your own adapters and and load them in

um but if you want to you know what we're hoping to do is encourage people to you know kind of submit adapters to our collection and we'll kind of include them is there a general skeleton that I can just adapt for different subgraphs or I have to build the adapter from scratch you have to build from scratch but again it's like you know I showed some of the code like the code is like like a simple adapter is like 50 lines of you know typescript so and you can look at all the existing ones and kind of figure out from there if you're just querying a um a graphq you know like the graph then it's a pretty simple adapter to to do that so thank you yeah no worries okay any more questions from the

audience okay then I have a question if I may go ahead uh you touch and they were jumping around a little bit here but you touched a little bit a little bit on kind of moving on from coin voting I know it's a tough question but like where do we move on to from coin voting what do you have like can you share some thoughts on that it's a good question and I certainly don't have the the answer um but you know in crypto we certainly like to like really reinvent things from a base level so I think the you know the end game of how like on-chain governance and not just governance but on-chain communities look is probably similar to the way organizations work in the real world right like corporations have coin voting you have share you know shareholders

that can vote on things but typically one thing they do is they they vote for the board the board kind of selects Executives Executives hire people like you have kind of this whole organization tree and so if I had to guess I'd say that's probably kind of the the end state for kind of digital organizations as well okay so like replicating existing traditional organizations but just on chain on chain governance yeah but with more more transparency more you know control like you know of course like it's really hard to to do shareholder voting it can happen like once a quarter things are not very transparent so we can certainly use the new tools we have available to try to um be more efficient and you know to to maybe do some things that wouldn't be possible um you know like liquidity mining is you

know that's that's not related to governance but that's something you can't do with stock shares but you can do with crypto but in terms of organizing people I think you know crypto doesn't really change you know the need for people to have some some level of organization right cool okay thank you very much David thank you guys [Applause] foreign