Epicenter - Learn about Crypto, Blockchain, Ethereum, Bitcoin and Distributed Technologies - Fredrik Haga: Dune Analytics – The Open Data Platform

Episode Date: November 30, 2022

Dune Analytics is a platform aggregating blockchain data, making data easier to query and understand for the larger public. Over the past years, Dune has become the go-to tool for blockchain analytics, as well as the home of a community of dune wizards numbering in the thousands, with its fundraising at a 1 Billion valuation as a testament to the growth of the project over the past 4 years.In today's episode, we are joined by the project's co-founder, Fredrik Haga. We discuss the products Dune Analytics offers, why it's hard to aggregate data that in principle is public, and how a community of data wizards was built in a grassroots effort. We also chatted about product focus and vision for the future, as well as useful lessons for bootstrapping a project on a shoestring budget.Topics covered in this episode:The beginning of DuneThe challenges and difficulties in building a blockchain data product.Technical improvements on Dune V2Supporting for non-evm chainsBusiness models, userbases, and valuationsInteresting use cases on DuneCommunity BuildingThe focus on product experience and next steps for the projectEpisode links:Dune Analytics's websiteDune Analytics TwitterFredrik Haga's twitterSponsors:Omni: Access all of Web3 in one easy-to-use wallet! Earn and manage assets at once with Omni's built-in staking, yield vaults, bridges, swaps and NFT support. https://omni.app/ -This episode is hosted by Friederike Ernst & Felix Lutsch. Show notes and listening options: epicenter.tv/472

This is Epicenter Episode 472 with guest Frederick Haga. Welcome to Epicenter, the show which talks about the technologies, projects and people driving decentralization and the blockchain revolution. I'm Felix and I'm here with Frederica Ernst today. Today we're speaking with Frederick Haga, who is the co-founder of Dune Analytics. Dune is an analytics platform focused on blockchain data. Before we talked with Frederick about Dune Analytics, we'd like to tell you about our sponsors this week. Omni is your new favorite multi-chain mobile world. Omni supports more than 25 protocols so you can manage all of your assets in one place across all major EVMs, layer 2's, ZK Sync, and StartNet coming soon, and non-EVMs.

But what's really special about Omni is that you can do all the most important things in Web 3 directly within the wallet itself. Want to get yield? Omni allows you to get the best APIs with zero fees in three tabs, be it staking, liquid staking, landing, or yield vaults. Need to exchange USDC on ETH to Atom on Cosmos. Omni aggregates all major bridges and Dexis so you can be able to be. bridge and swap across all supported networks in one transaction. Now of NFTs, Omnis offers the broadest NFT support of any wallet, so you can collect and manage your favorite NFTs across all chains in one place. Omni truly is the best and easiest way to use Web3, and most importantly, it's fully self-custodial, meaning you never have to trust anyone with your assets other than yourself.

If you want to use, you can even use Omni's ledger integration so all your funds stay on your hardware wallet. Join tens of thousands of this next generation wallet by downloading it today on iOS or Android at Omni.app. So, welcome Frederick to Epicenter. Thank you so much. Very excited to be here. Epicenter was one of my first main sources

for understanding this crypto world when I got excited about it back in 2015. So cool to be here. Yeah, awesome. I think we actually met at a epicenter meetup in DefCon4 in Prague. So really awesome to have you on today. And obviously you come a long way.

I think that's also where we wanted to start, right? Like basically how did you come up with the idea for Dune? Kind of like can explain a little bit what is you and how did you get into crypto and to build Dune? For sure. Yeah. So me and my co-founder, we were working at a big company back in 2017 when the sort of hype

happened for the ICAO crazy and all that. And we, I was working full time. on blockchain, I was already interested and managed to sort of work on it because of all the interest. And then I met my co-founder. He was also interested. And we ended up forming a small team, doing a lot of prototypes and playing around with blockchain and building on Ethereum first and foremost and built prototypes and did hackathons and all that. And then over time, we realized that during the spring of 2018, that the hype is kind of dying and people, people are not that interested anymore and this technology is still very young and clunky.

And so like this big corporation might not be that interested anymore, which was fine and understandable, honestly. So, but we were never the less just as excited about the technology and the potential for, you know, open finance and programmable assets on the internet and whatnot. So we started thinking about problems to be solved really in the space. And we had, since we had built and deployed some smart contract, we knew that getting the data was was pretty clunky and hard. And so that's kind of what seeded the idea for starting June.

And then over time, we also understood that the fact that this data is inherently open is like a total game changer for how you can build a data product in this space. And, you know, it's a huge paradigm shift, whereas previously all data and all data tools is like in close environments, like inside the engineering or product organization building Web 2 product or inside of a financial institution. And the name of the game is really closed and, you know, building tools around proprietary data sets and whatnot. And so I think that the thing that we understood pretty early on was that, oh, when the data

is open, like you can actually aggregate, you know, aggregate all of this knowledge. and building that people do in the space as opposed to defaulting to private. And we sort of flipped it on its head and said it's default public, right? So that's kind of how we got started. And in the beginning it was like really hard, impossible to raise any funds. We didn't have a salary for many months. And yeah, we traveled to a few of these like eat hackitons and devcon and whatnot on like a very tiny budget sharing hotel bedrooms, Mots and me and just try to find anyone that

was interested. And it was hard because not that many products were actually live with the usable products at the time, right? So the sort of market was small, but there were a dedicated group of people that kept building and were excited. And those were the ones we tried to serve. And I think that has turned out to be a good idea to bet on the people also betting on this future. So in a way, what you've done is kind of you've, aggregated an enormous database that people can easily carry. And basically, your products kind of fall into two divisions, so the dashboards and the API. Can you kind of talk about the distinction between those and what exactly they offer?

Sure. So in its simplest form, what we do is we take data from the blockchain nodes, we put it in a database, We do some transformations to make it easier to work with. So you have smart contract events. You have other abstractions on top of that as well. So you can see, oh, this is a uniswob trade. But you can also say, oh, here's a table with all the trades from all the dexes and, you know, easily query that. So we put a lot of effort into good sort of data UX.

And what we do then is to allow anyone to go to our website, browse all these takes. labels, write SQL queries on top of it, get results, and then visualize it and turn it into dashboards. And I guess what I alluded to earlier is the fact that this is all public by default. So if you go on Dune and you do something, someone else can see it, they can see the sequel, they can fork it, they can make their own version and whatnot. So it's all kind of remixable and open for anyone to build on top of each other. And then also we're about to launch and an API product, which has been highly demanded for a long time, and essentially take any of these endpoints that underpin these dashboards and query results

and turn it into programmatically accessible API endpoints, which I think is going to be a massive enabler for a lot of people to not have to do all of this tedious data infrastructure and maintenance of pipelines and whatnot to query this data. Yeah, we're going to get a bit into like what the use cases are. I think one question that maybe people are interested in is like if you talk about the dashboards, obviously we have a lot on Dune. Do you have a favorite dashboard yourself and then which one is it? So I've had a favorite for a while and I think it's maybe even more fashionable right now,

which is MakerDAO's balance sheet built on Dune. And I think this was a particularly accessible. exciting example. I think they built it out a couple of years ago, really. But the fact that you can have a real time balance sheet feeding from the blockchain that anyone in the world can go and look at on a website and they can see which query created this balance sheet. And, you know, if you don't like it, you can fork it and you can massage it, right? So it's really very scrutinizable. And the fact that it's real time and you can see which assets, which liabilities, all these things. and the revenue, right, block for block, you can see the revenue coming in.

And so you can do like all the way from like financial auditing to even investing. And you can say, oh, I can have a real time price to earnings ratio for the maker token, right? You can look at that metric that analysts in the traditional equity world would figure out once, four times a year, whenever the PDF drops, you could say, oh, now it's priced this relative to the last quarter earnings, right? But like literally you can do that every 12 seconds on the Ethereum blockchain with the Dune dashboard. And I think this was very illuminating because it displayed, you know, people have been looking at all kinds of weird metrics on Dune, but this displayed, you know, the old world, kind of, but in this new world of ours on blockchains. And I think that was a very illuminating example. And then, of course, with the FDX poll out recently and their cooking of books and whatnot or lack of books altogether maybe. It illustrates how insane it is that MakerDAO and whatever other financial product built on the blockchain is there for people to audit, to monitor real time.

And there's no hiding. It's like completely illuminated. And Ethereum has been called this like dark forest. But I think like, Dune is a huge flashlight into that dark forest, illuminating everything you want to see really. you can get there and you can look at it. And I think the beauty of June is that this is not, you don't have to be deeply technical to go there. You can look at what other people are doing and you can tinker around.

And you don't need to set up data pipelines and infrastructure to understand, which I think is powerful. The dark pool leg is usually used to to refer to the mempool though, right? It's not, I mean, so basically everything that happens on blockchain is by default, open. I'm not going to say transparent because I totally hear you that basically getting the data is not always easy. It is possible though. But maybe let's talk about the difficulties that you guys have navigated, the challenges that you have grown to with June. So basically, why is actually getting this data together? Why is it difficult?

Yeah, no, it's a great question. So I think, well, so a blockchain is a system that's built for like verifying transactions and, you know, activity as it happens and letting people sort of right to the blockchain, right? But it's not necessarily optimized for getting data out of it in large quantities. right so so already building a blockchain is a hard engineering problem and the constraints that you would focus on is the verification how many can easily you know verify it and then you can actually make transactions and interact with it so the whole sort of let's call it big data aspect of it is not really what that system is designed for and so that means that it's actually surprisingly complicated to get the data out.

And you need to go block for block and ask for all the different things that happens in the chain. And since like a no, there's a piece of software that's not optimized for this, this is way more messy than most people think. And it's not like you just hit play and all the data just arrives. Like things happen. There's weird transactions that stole the system. And I think the most extreme example or clear example of these things are these sort of get forks, so something like Binan smart chain or polygon that have tweaked the parameters of the get node with essentially lower block time and higher throughput, right? And so what happens then is that some of these nodes have a really hard time sort of reproducing. something that happened, say, you know, a million blocks ago.

And so it means that if their block time is three seconds, but they might spend, so spend like a minute producing a trace of something that happened. And that means when you're backfilling that chain, like you diverge from head. If it takes one minute to get a block and is produced, you know, 30 or 20 blocks. in that time, you've gotten one more block and then you're 20 blocks behind, right? Further behind than what you were before. So since these are systems that are not really built for taking the data out, you do have the issue of like all these kinds of bottlenecks and quirks.

And so it requires a lot of handholding to just get the data out. And then of course, just like having a performant infrastructure and system in place to efficiently query what's becoming a lot of data. And, you know, there's Ethereum, but there's also these other EVM chains that typically have more throughput and produces more data. And then you have things like Solana, which kind of thinking in less than a week produce more data than the whole of Ethereum's history. So, you know, there are non-trivial amounts of data as well to query.

Yeah, I totally get that. Maybe let's dig down a little bit deeper, though. Just, I mean, maybe let's stay on Ethereum exclusively, because I totally understand that basically other chains with, you know, different block gas limits and so on have higher amounts. Let's say I have an archival node running. And that basically gives me the state of each variable at each block, right? So why would it be difficult for me to kind of run?

run a query on that archival note. I mean, I mean, running an archival note itself is a non-trivial thing, right? But I mean, if, if you're going to do that, what exactly is the bottleneck for me if I want to grab like say the unyswap volume for a specific token pair over the last two years? Yeah. So, you know, I think you need to know then exactly what you, exactly what you want to look at, which is also not trivial, right? So you say, okay, I want to look at Unispop volume for the last few years.

Okay, that's, I don't know the exact number, but it's probably at least in the tens of thousands, maybe in the hundreds of thousands of contracts. So you first need to figure out, like, which contracts do I care about, which are definitely a lot because they deploy a new contract for every market that they create or every peer, right? So you need to somehow know what data you're looking for from the node. And then, of course, you get that. Then that in and of itself won't give it the answer because you also, for instance, would need a USD price. Like if you want to aggregate all this volume, you need to know, okay, if like you did, you know, cat coin and doge coin are traded, what are those actually worth?

That's, you know, some data you would get from the blockchain. but wouldn't be practically that useful to need something else. So you need the USD volumes. You need to fetch that somewhere and join it and, you know, multiply the amounts. So you have that type of problem. And then you have other sort of higher level sort of curation problems where there's a lot of spam, for instance. So someone will create some nonsensical token, you know, do one trade that makes it worth a billion. And then you can do a massive trade that looks like it's like one billion dollars worth of volume,

but it turns out it's like, you know, just some folks trading with each other on some worthless token, right? So there's also like that element to it. So I think it's like both technically, non-trivial. There's also a lot of context that you need to do this. So basically it's the integration of different sorts of information into one database together with kind of the categorization and curation of different contracts and feeds. Is that high level correct? Yeah, I'd say so. You know, also sort of the benefit of what we do is that we fetch absolutely all of the data, right? And so you can sort of have that as a starting point. But if you interact with the no, you know, if you do want to answer some specific question, you do need to know what data you want to do,

but then, and you can run that, and it's going to take you a long, long time, and you have to sort of handhold the process. And then if you say, oh, but actually I want to check this other thing or match it with this other thing, you know, then you have to do that whole exercise over again, where it's like, oh, but actually, you know, I want the question related to balancer pools, you know, then you have like a whole process going. And so, yeah, I think it's essentially just in theory doable, but in practice, like a lot of hard work and a lot of mess, even for a sophisticated sort of data engineer. Right. And you recently also upgraded, I think maybe going a bit to the technical part of it from like V1 due into V2. Can you talk a little bit about what you've changed there and how it improves the Dune experience? Or also like what was the problems with the Dune experience so far?

Yeah, for sure. So we had a Postgres-based system for Dune, we won, which was a great start for a Dune, a great piece of technology, but with a lot of data, it became a problem. So basically, a lot of queries would time out. and like a post-coast type of system is good at sort of being fast on relatively small amounts of data, but not great when it comes to a lot of data. And so there were a couple of limitations in addition to performance over large data set. Like you couldn't look at, say, like median gas price over the history of the Ethereum blockchain.

We just timeout. We run for like 30 plus minutes and then, you know, you wouldn't get a result. There was also the problem that each chain we added was a different database and you couldn't work across them. So you couldn't benchmark data on polygon and Ethereum. They had to be like separate queries. And then furthermore, like all the data pipelines of creating these tables that we have like Dext trades and NFT trades and these abstractions I talked about was like very messy because these were like scripts that pulled some data and sort of referencing each other, like, very, very complicated to maintain and understand which data relies on which data, on what tables and whatnot. So essentially, we've now launched through June Engine v2, which is based on like a data lake and more like a modern cloud data warehouse architecture, where we have all the data for all the different blockchain.

in like a data lake and then we have a query engine on top of it. Currently, it's SparkBased, but we're also doing some exciting new things there. And essentially allowing you to traverse all the different sources. So you can benchmark Solana and Ethereum and Polygon and all these things in one simple query. It allows us to have way better pipelines and checks and balances on how the tables are created. Also, we can now have commune. community contributions on the data layer. So not only can people write queries,

they can also create new tables with something with a product we called spellbook. So people can then create all kinds of derived data sets that are easy to query. They can write their own tests and ensure that the data is of high quality. And so it's a system that's way easier to maintain. We can invite the community into, contributing more and sort of self-serving to more use cases. And then, yeah, there are some amazing performance benefits and sort of what the typical modern data stack looks like. We're kind of, you know, creating that and enabling that

for the community where you can do median gas price in, you know, a couple of seconds or a few minutes at least. You just alluded to the fact that you now also support Solan. which is a new development. So previously it was only EVM chains. And despite the fact that kind of the block gas limits and block times kind of vary between different EVM chains, I assume the basic infrastructures fairly similar between all of these. How are things different for Solana? And how did you make the decision to support Solana over a number of other non-EVM layer one projects

that also have traction like Cosmos, or substrate or avalanche or similar. Yeah, great question. So, so first of all, like, we want to support all blockchains and everything that's interesting. So it's not the goal in and of itself to limit it. I think what we have cared a lot about is the user experience of, you know, using the data on June. And I think that's frankly one of the things we got right were, You know, the easiest way to make some money being a data business in like 2018 or 2019

was to like just integrate with all the chains that raised a ton of money. We didn't do that. Like we stuck to Ethereum where essentially no one would pay us to do this. But we figured, you know, this is where people are building. So let's serve them. And so for us, the important thing is that there's an ecosystem of developers, of applications, of users. and communities that actually want to query this data and, you know, do interesting things with it. And then sort of we do the work of the integrations.

And as you alluded to, it's been way easy for us with EVM chains because all the infrastructure, all the processes are the same. So when we add something like polygon, like it is, there is definitely challenges with every system one learns, but, you know, in broad strokes, a lot of the architecture and processes, processes are the same. So with something like Sulana, you know, this was actually the first also blockchain we put on our new Junev2 engine. So we figured this was a nice stress test to the system because it they produce a lot of data. But we saw that there was a lot of people building on it. There were actual products being built, users using them. And then we decided to integrate

it. And we honestly also have built out this new platform, which has been also a learning experience and something we've had to work a lot on and therefore have been painfully slow, honestly, in integrating some of these new blockchains. But now the new platform is getting into a more stable state, and then we can go more aggressively on integrating new chains. So it's not something we deliberately sort of try to avoid, but we try to make sure that the UX, when we do integrate, is good and that there's an interested community to work on top of it.

And of course, like the technical complexity plays in if it's a entirely new system, then it is more work to get it right. Cool. Yeah, we wanted to talk a little bit about it kind of Dune's business model and also, I guess, the user base. You mentioned already, like the community and everything. I think one interesting thing about June is that it's basically one of the few products that has this

kind of SaaS model, let's say, in the in the Web 3 world. And we were interested to hear a bit about, you know, like also how many users are there on Dune, how do you think about surprising, how much revenues are there, whatever you can tell us about kind of how this all works. We'd be happy to learn. Where do I start? I think we have built a very strong community on Dune. I think we've actually, and without the token, which I think actually has been a good thing for us, frankly, because there is like a kind of proof of work on getting into Dune.

You do have to make some sacrifice, put in some work. And that has made our community very sticky and healthy, I think, because people actually put in work to get there and learn. I think we empower a lot of people. So as I alluded to earlier, like, it's not only developers that use Dune, but a lot of non-technical people learn SQL with them for Dune. They do some tutorials, they fork some queries, and they get started, and suddenly they get paid to work on Dune.

They pick up bounties. They get a job. A lot of people have claimed, like, fantastic, top jobs in protocols. and venture firms and whatnot, starting from basically nothing and having no online presence and then working on Dune, building some nice important dashboards and getting like top jobs, right? So I think that that's very important part of it, that I often think of sort of our community and sort of the Dune team building the product in terms of this Spider-Man meme where they're

like, oh, you know, you built a great tool, you helped me do this. and we're like, no, you did a lot of cool queries and you helped us. So I think there's a lot of mutual appreciation between the people building Dune and our community, which I think is a fantastic thing. So in terms of how big it is, we have like a few thousand people creating queries a month. We don't have like a huge user base on the creator side.

But since Dune is kind of like, like a marketplace all the way, almost where people can come in and see anything that's built, it's like a creator side and a consumer side. And so we believe that if we empower creators, they can make amazing things, and a lot of people can benefit from that. So we don't necessarily need 100,000 query creators. If we have someone that's very engaged and create amazing things, we do have like millions of page views.

And then some of those convert to creators learn the craft and, you know, get into it. So that's kind of the dynamic. We have about 50,000 dashboards on June underpinned by like 300,000 queries. So it's quite substantial. Yeah. So I think the heart of the question goes to this. So currently you operate under this, you know, description model, right? So basically, there's different tiers. So basically, you have to pay for

kind of keeping your dashboards private and having the watermark removed and so on. And then this kind of, when you compare this with your recent raises, congratulations, by the way. So you've raised repeatedly over the last couple of years. So at first it was $2 million and $8 million and recently $70 million at a valuation of $1 billion. So my question goes to this. So the investment that's been made, was that an equity investment or was it a soft? I mean, basically to kind of reframe this question. Is there a plan to introduce the token? Because basically, I struggled to kind of make sense out of the software as a service business model that you're currently running with these variations because I mean sure there's kind of companies for this for for for whom that works like

the sales sales forces of the world but obviously your um your your your addressable market is so much smaller so how does this how does this all fit together yeah sure um so so let's start with like the business model which i which i didn't intend to to skip over from from the previous question So essentially, you know, what we do is we allow anyone to use this tool for free. We swallow all the infrastructure costs, all the work in terms of making this product super useful. And then our philosophy is if you do it in the open, anyone else can leverage what you do. You build sort of in public. And so that's free.

What we essentially charge for is if you want to keep things private, you want to do it only for yourself. if you want to use more resources, so have more performance on your queries, have more beefy infrastructure underneath it, have shorter wait times, these type of things. And if you want to export data, which essentially mean that you can do proprietary research, build your own business, whatever it might be, right? So these are kind of the three main dimensions that we think make sense to monetize along. And I think, you know, to back up a bit, there are different ways to build a system and sort of say, you know, what do we kind of give away and what do we maintain to make a business or whatever? And I think we've kind of our approach is like we want to solve problems, we want to be as useful as possible.

And what we've figured out is like if we do it in this way, we can have the most impact because so many people, so like, yeah, sure, we could open source everything and just like everything that we ever did is open source. But that actually would have way less impact. It's our hypothesis because so many people wouldn't be able to interface with these nodes to actually run this data infrastructure to operate and maintain this thing. You know, we have almost 50 engineers working full time on this. It's not a trivial problem. And so, you know, that's the model we've taken where we think, you know, the community can get a lot from this and we can monetize the people that have extra needs in different ways, but also sort of, you know, have more of a business use case typically. rather than just contributing.

So that's kind of the philosophy. We've raised equity. It's not that I want to completely rule out any other sort of Web3 play. But I think for us, we're very concerned with taking long-term solutions. I want to, you know, if we wanted to get rich as quickly as possible, like we would have done due it in a different way. But we care about building something that, you know, it's long-term that is sensible, even though prices are up 100% or down 90%, like, I don't really care.

And so for now, we haven't found a model where we can swallow, you know, massive infrastructure builds and also kind of have a different structure. But over time, we definitely, you know, try to give economic opportunities to our creators as much as possible wizards building on dune. And so we'll always be mindful of the toolkit and what we can do. But I think it's also from our side, like ensuring that we don't get ourselves into a weird spot where we fall prey to some hype and then suddenly like, you know, six months later, you're in a really bad spot because you did something that actually doesn't make that much sense. And also like just pragmatically solving problems every single day and ensuring that we're

massively building something useful for for the community. And you are. So, Frederick, I've used your dashboards and I've built my own dashboards. And Dune is such a fantastic tool. And I mean this in the least adverse. I know this sounds a little bit adversarial. But basically, so I'm a little bit struggling to understand where your mode comes from. Right.

So basically, even if you, I mean, you have 50 engineers. And maybe we'll talk about what they do on a day-to-day basis in just a little bit. But say, you're paying your engineers a really good. good salary, say 200k here, that's 10 million a year for 50 engineers, right? And then maybe add the same amount again for your AWS bill, which is probably around the right number, I'm guessing. So it's 20 million a year to recreate something like June, but you still have this enormous valuation. So basically, where's this expectation of, profit from the VCs coming from.

Yeah, no, great question. I'm not going to comment particularly on your estimates of how much it costs to run Dune, but fair enough. I think, you know, I think we're at the end of the day, we're trying to build something useful and I think we're abstracting away a lot of hard work for a lot of people. And I think if you run the protocol or a VC or, you know, other type of players, like we have big, you know, big brand names and whatnot playing in the space now. You care about this data.

Like, it's important for your investing, for your running of your business and whatnot. And I think, you know, the things I mentioned, like having more performance on working with this data, having a private environment where your team can collaborate with this data, exporting this data through API and these type of things. I think it's like massively useful and people will pay for that convenience. I think that's valuable and the fact that people can come together. And as I mentioned, like, you know, we get a lot of people that contribute and we, the tradeoff we kind of made is like we say, okay, we will swallow all these bills. We will make this system run. We will ensure that like, you know, everything is up and

running and correct, which is a massive effort. And then I think there's, you know, a lot of demand for getting more out of that tool. And I don't think that is necessarily, like we care deeply about the free experience on June. And so it's not like we will suddenly just like rug all the free users and, you know, extract massive amounts of money from them. It's like we will make June more powerful over time and let people pay for extra features.

And I think it's important sort of context here as well. that like we did a play on like the long time when we focused on the community in 2018. And yes, like I didn't have a salary for seven months because every single VC was like, there are five people you can sell this to. And we don't see this as a lucrative business. And we were like, actually, we think this five people is going to become like thousand and then thousands and whatnot, you know, over it over the years. And that's still the play we're doing.

Yeah. Don't get me wrong. I think Junice is a fantastically useful to it, totally. How do you feel about the comparison that's sometimes made with Bloomberg Tamara? I think that's understatement of what Junice can be. So I think, as I alluded to earlier, like the fact that this data is open is like a total, total game changer.

And Bloomberg is a fantastic product for a new, niche in the finance world that, you know, care deeply about having a sort of proprietary understanding of whatever is happening in these financial markets. But and so, you know, amazing product and fantastic what they've done and the position they have in that world. I think the fact that you can have open-ended interaction with this data as like together with the rest of the world is just. enormous. And so I don't think of like I really don't think there's any product out there that that is comparable to what we're doing because it has some of the aspects of like a GitHub where a lot of people come together and build on each other and and you know, it's a tool for for developers. And then you have the product of that which GitHub doesn't have but which kind of, you know, it's more of the Bloomberg thing. like, oh, there's actually a bunch of charts and a lot of people can just look at, which is incredibly useful.

But then you also have, you know, the fact that this is public, and which I think is a huge deal. And I think big kind of technology paradigms, typically what you see to succeed is like someone that really builds for some novel new aspect of a new platform or a new piece of technology. And to me, that is really what we're doing. here. We're not trying to recreate Bloomberg or recreate some tool that people had for their product analytics and they would like pipe their whatever sort of usage data from their website into a tool like that. Like we're trying to build something that is uniquely enabled by and driving open data and understanding and utility of that data. And just the fact that when there's a trade on on like a balancer pool,

previously, like, there would be, like, one or two stakeholders to that data point would be, like, the PM and the leader of that company that built that product. Like, now there's, like, this explosion of stakeholders where it's, like, literally anyone can go on the internet and look into this data point and, like, make sense of it in whatever shape or form they want to.

And I think it's still hard to grasp and appreciate, like, how big of a deal that is. that like the world is the audience here. And of course, not every single person in the world will go there. But it's an opportunity. And I think that is the premise of what we're trying to do is like people can learn faster if they see what other products are built and they can see the open source smart contracts,

but they can also see what traction that smart contract got. How do people use it? How do people abuse it? You know, and I think what the blockchain is, one of the exciting things about the blockchain is like anyone can build, anyone can participate. And you have this sort of bet on human creativity that like you just give people tools to do whatever they, you know, there will be a lot of stupid ideas, but there will be some really good ones. And I think what we're enabling in that is that you can actually dissect what worked, not just with the thing you built yourself, but with everything that anyone. ever built. And then you have this amazing sort of feedback loop and evolutionary force that's just so much faster and so much, you know, it's harsh also because if you build a product, like,

it's out there. Like, anyone can go and look at your metrics, not just your manager, which is, of course, intimidating, but it's also like, I think progressing this space way, way faster than any other industry that you can think of. Yeah, awesome. I want to dive a little bit deeper into that. So we we had mentioned a little bit right the maker dow balance sheet we have also mentioned kind of other dashboards that have been built i would be curious to hear from you you know what are there use cases you found like really interesting or even like people building on top of dune in a way that maybe maybe you didn't expect or also if that hasn't happened yet what what you would like to see build on dune i i guess there there is kind of like two different things one one is uh

or two different worlds. One is when we launched the API, which is going live. By the end of the air, for self-serve, we have kind of like a waitlist, private beta right now. You know, that's going to be an explosion of opportunity. But inside of the Dune app, I think, frankly, I do think that in many cases, the crypto community should learn. a little bit more from the sort of old world. I think a lot about, you know, what are features and what are bugs of the old world. And I think still too many crypto people think there's only bugs, you know,

in the old world and everything needs to be reinvented. And I think maybe, you know, there surely are some bugs. And I think we have an interesting new design space. But there might be some features after, you know, industries that have been around for a while. And I still think kind of startup one-on-one metrics, are underrated. What's the traction of your product? What's the user base? What's the retention of that user base? You know, what's the distribution of activity across different users?

And I still think what like a typical startup PM would want to look at is actually not that far from what people should look at in crypto as well. And like, do the boring stuff. Just like figure out, are we retaining users? Like, will this person come back or not? And if they don't come back, like, why? So I think that is still a bit under appreciated and people like to look at sort of TVL and some high level like trading volume, which is probably a better metric than TVL at least. But I think that some of those kind of more sobering metrics is something that would be nice to look at.

there was also like a very cool longer form research piece by a committee member last week on just like the uni air drop and you know what did that turn into and you know those are some pretty sobering metrics in there where it's like okay i can't remember the exact metrics but it's something like 90% sole more or less instantly it's like 7% of all the airbusy drop holders still hold the token and like a tiny fraction of that sort of increased their exposure and even smaller portion participate in governance right only only like a tiny tiny bit and is the unispop volume driven by people that got the airdrop no like that also fell from like 60 a year ago or something to like five percent now right so

So it didn't retain users. Most people just saw it as free money and dumped it. And it didn't actually turn into like strong governance participation. Right. And so I think that's those type of metrics and lessons, right, where people should say, what's the return on this initiative that I did? Like what did we try to achieve and what did it turn into? It's like an exercise that more people in crypto.

should do, I think, and be honest with themselves. And I think obviously, like these times when prices are down, the sentiment is down, like you kind of have to do these things. And they matter, I guess, less when everything is up anyway. So I feel like, you know, in general, like being a bit hard on yourself and thinking about the stickiness of your product, the ROI of your activities. And in that post, maybe we can link to it. There are other aerobes as well that you can look at the metrics and just see for yourself.

Like, you know, did they actually engage a community by doing this or did they not? And if they didn't, like it's an incredibly expensive, you know, one-off marketing gig that kind of, you know, I can get a lot of people to use my product if I just go out and say, hey, here's like $1,000 on the street, right? It doesn't necessarily mean that I'm building a fantastic product or business. So, yeah, I think it's good timing for just like reading up on some startup one and one and start to think about the ROI of your initiatives. So user experience and a good, hard look in the mirror. I mean, clearly that's super advantageous for projects. Do you think there's something for the ecosystem? that the ecosystem can use to its advantage.

Because obviously, I mean, volatility in crypto is large, despite the fact that in principle, a lot of the data is public. So, I mean, obviously there's some data that's not looked at or that kind of is subject to externalities. Do you think we can kind of leverage June to that end as well? I guess you can definitely learn a lot from history, right? And I think it's probably a good, now we've had a lot of lessons. And the pace has been incredibly high.

You've had prices going up and prices crashing down. And you have all these systems that's been built and all these things that happen, air drops and vampire attacks and whatnot. So I think it's probably a good time to be a student of the on-chain activity. Right. And maybe you don't even have to be an engineer to do that, right? But if you're looking for products to build, I think it's like an amazing time for learning some lessons.

And there's been a lot of stress tests. And I think now instead of just the end, you know, the FIRE NFT project being thrown together in the week and then raising $10 million in two weeks, you know, people can actually sit down and say, okay, you know, did any of this actually work? Are there interesting pieces here? So I don't know exactly what to look at or not, frankly, you know, but I'd be mindful of especially this like, you know, what I think one thing that's underrated a bit in crypto is like if someone is getting paid for something, someone is picking up the bill in one way or another. Like there are no free lunches. Like if you have some kind of like token emission scheme related to using your product, like the other token holders are taking dilution. there's no way ever that money just like appears out of thin air and so I think that is at least something to be very mindful of that you know the way you design these things someone's going to pay a price if other people are receiving money so you know think about that yeah absolutely you have been extraordinarily successful in kind of building this community of Dune visits, as you call them.

I would actually struggle to name one infrastructure project that kind of has a better community than you guys. How did you do that? Do it. And what are your learnings? Well, thank you for that. So for the first two years, Dune was just Motsamie.

Like the only, we hired no one. And as I mentioned, you know, had no salary and then very little salary. and we didn't hire anyone after our first fundraise or anything like that. So we're very lean. But I think we had a very, very deep passion for helping people on due and making them succeed. Like our telegrams and Twitter DMs and like we, there weren't that many. Like there were pretty few users, but we were extremely committed to making them succeed.

And we did very hands-on support for helping people succeed. And this includes several people that didn't know. SQL, but we like help them learn sequels so that they could use their due, right? And and for instance, Tio from then researcher at the block and then now at Uniswap, he, he was like, hey, this seems like a cool product. Can I get it in Excel? And I was like, no, but I can learn you how to do SQL. And we had like a biweekly call where I would like, you know, give him lessons in how to to use dude. So that was the level of commitment. And I think that has been. you know, kind of snowballed into what is the Dune community because, you know, our first community

manager, Roxxer, he had that similar passion. And I think every single person that has come into the Dune community and felt this support and people saying, yeah, let's get on a call, you know, let DM into the late hours on like trying to make this query run. I think that has really, you know, compounded to a culture where people will say, oh, yeah, you know, let me help you out. And I think one of the things I'm most proud of with everything we built is really the Dune community and just like going to the Discord, during this course, like several thousand people. But it's still like mostly just people asking about queries and getting answered to those queries. It's not a ton of crap.

And, you know, I think it's still very much active even though prices are down 90%. And I think that comes from. a passion for really helping people and being as, you know, there was this funny thing where like Mott's, my co-founder, he recently realized he had like all his telegram notifications turned off, but he had like an exception for one of our early users. So that would like get straight to the home screen of his phone. Like if they DM'd him on telegram, like it would pop up on his phone on Friday night and he would like get in there and make that query run. So, so I think that passion that we put into it is something that has actually scaled into a wider community.

Yeah, super interesting. So, so you're saying you didn't have to give $1,000 to every Dune wizard to use the product. Exactly not. Right. Awesome. Yeah, I think maybe I guess are you working on like specific things? We know there was also the DuneCon. Maybe you can also talk a little bit about their first conference but I think from what I hear it's a lot yeah it on the field like helping in discord but do you also have certain I don't know learning material like documentation is that something you're working on or how do you improve that even even further yeah no definitely we're working on a lot of educational efforts we have a big community team now we're doing meetups and yeah we did dune con which was a massive success in in Berlin a couple

months ago. So we really tried to get people engaged and help them out, both sort of hands-on, but definitely also more self-serve with the educational series and whatnot. So we are working on several sort of bigger initiatives. It is one thing that's, you know, honestly a challenge is like we were doing this like system migrations and like we have the V1 and then we have the V2 and like there's a different dialect of SQL and like, you know, these are kind of some of the the messy things of building a company where it's like, oh, yeah, we want to have all the best educational resources, obviously, but then, you know, oh, there are different versions and which one, you know, can only support and what features do we have today, then what do we

have tomorrow and whatnot. But yeah, we're certainly keeping investing in education and letting people become powerful wizards and sort of leveling up and giving them bounty opportunities. And we have a job board and all these type of things. And we have seen several people sort of pivot their careers into just doing this, which is fantastic. And then, yeah, there are some, you know, we touched upon it, but the Dune API, I think, is huge, huge game changer. Like what you can do with data in this space is going to be forever changed when you can

just click a button and turn any query salt on the unit to an endpoint that you can leverage for building or trading or researching or, you know, whatever you see fit. So I'm pretty pumped for that. So SQL is pretty straightforward in terms of, you know, programming languages or databases. So if you know two things about databases, you can carry it. But still, there's a lot of people who are kind of afraid of, code, right? And in principle, it should be relatively easy or comparatively easy when compared with other programming instances to kind of open this to natural language processing,

kind of like what from Alpha has done for Mathematica, right? Have you looked into that? Because it would immediately balloon your target audience, you know, from a couple of thousand people to, literally millions. Yeah, great, great question. I think the big challenge there is kind of going from, I think it's like relatively easy to build a cool prototype that, you know, takes some of the like most obvious things and turns it into like a result and like creates a SQL query for you underneath the hood. So I think, you know, that is very cool.

I guess what I'm wondering is, you know, to what extent does that extend to like all kinds of weird, messy things that humans, you know, would actually put into like a search box and how could use, to what extent can you serve that into a compelling product experience? And how many people will actually go and do that versus just like look at something that's kind of pre-curated? it. So I think for us, like, we're very mindful of just trying to do like, you know, the most obvious thing first. And we feel like still there's a lot of work to just like make the querying experience on Dune better, have cleaner tables, have it run faster, some of these type of things that are just going to make writing SQL on Dune even easier, even faster and more powerful. And then over time, like, we're excited about. other ways of interacting with this data, if that's Python and notebooks or, you know,

some natural language sort of experience. But I think it's still, you know, we still think that we have some more basic things to get to. And I'm also, I think, yeah, one need to be very conscious of will people actually use this product and like how good will it be. And especially if you're thinking more of like a consumer. audience, it needs to be very polished. And I don't think you can have that many rough edges before people sort of just get

annoyed and don't use it. So yeah, yeah, fair enough. In a way, basically, if you look at you in trends, I actually do this regularly, at least every other day or so, just because it's super interesting to kind of see what other people find interesting inquiry, right? So in a way, you kind of get a front row seat to developments in the the ecosystem. So what's your take on where we're at? And what should we collectively pay more attention to? Yeah. No, I definitely have a hope that people can be a bit less on Twitter and be a

bit more on trending dashboards on Dune and we can get even better at surfacing, like what you should actually look at and that hack can just as well be discovered by looking at trending dashboards on Dune as on Twitter per se. But, yeah, I think a good thing to think about now is kind of risks. You know, there's all these pullouts. Obviously, a lot of this was actually not on chain products, but sort of things, just like in the industry that was off-chain. But I think understanding, you know, the risks, I think is a good idea.

And I guess people are now second guessing, so everything and all their exposure and all kinds of different products and what mechanisms kind of actually makes sense. So I feel like, you know, again, this is one of the powerful things of June. It's like instead of sticking your money into a black box and hoping that someone is taken care of it, like you can actually do your own research in terms of what that box is and what's in it and what the exposure might be or not. So I feel like, you know, some of these dissecting, like, as I mentioned, MakerDAO have actually done really good job at exposing their balance sheet, but that type of exercise for more products, I think makes sense what's in there. If you're getting some interest rate, you know, what's the reason for that? You can actually see what kind of exposure sits on the other end. Right. I guess one thing that we also shortly mentioned at the start was. basically FDX and that fallout, as you mentioned.

So I guess one question is, obviously Dune focused on on-chain data. Do you think Dune can also help these kind of centralized parties? Or do we need to move everything on chain? What's your take on that? Yeah, I think we can definitely help. We have a relatively new, like, labels data set that anyone can, contribute to. And I'd say, you know, we have some work to do to maybe streamline this and make it easier for the exchanges to actually show, you know, what their addresses are. But my sense

is that, you know, the actually legitimate exchanges now seem, you know, they care a lot about the confidence that people have in their business. So they are quite keen to show, you know, what their exposures are and what their on-chain activity looks like. And definitely, I think, you know, we have helped surface some of that. And the Dune community, you know, one thing I find amazing is like, whenever something happens, there is instantly dashboards, you know, right away that track what's happening on chain related to whatever event, including FDX and Alameda.

But I think we have a job to do to make it easier to sort of have, the exchange is actually, you know, verify this or say, you know, these are our addresses and easier instead of just like sort of submitting a blog post or whatever with like a list, you know, actually make it into an experience where a user can say, oh, you know, this is what they said their addresses are and here's the related activity on chain, right? Yeah, but it can only ever be like a part of the puzzle, right? because there's no way of really tracking liabilities on chain. I mean, obviously some liabilities you could, but, you know.

No, no, that's true. That's true. Of course, like, you know, and I think there will always be off-chain lending and leverage and agreements and whatnot. So we can, of course, only track what's on chain. But of course, the on-chain products, there you can actually see the exposure. So I guess the solution is spoken. more info from from the off chain businesses but also hopefully a bit more on chain uh usage as well

yeah so frederick what's what's next for dune uh you have 50 people working on stuff so what can we expect yes uh you can expect the dune API which uh is a very powerful um thing i think um and you can expect we are launching more team functionality so people can more easily collaborate, both in the private and public setting, so you can get into it with your team. We're working on more ways for people to create datasets and contribute, you know, make it easier to cost spells, as we say, in the spellbook product, create more of their own data tables. We're working on some very, very exciting sort of ways to make this interacting with this data more performant.

And sort of we are investing very deeply in the querying experience on Dune, helping both in sort of the interface layer, but also in the data and query engine layer to make it the most performance experience out there. That's one of the benefits of the scale we have is that we, you know, you asked earlier like what do the people do? And part of the answer is like we're really something that's not announced yet even, but like we're going very deep into the stack and customizing it for the world of blockchains. And I think it's a unique thing that we have this predefined data set that's like somewhat structured. And we want to build something that's like uniquely tailored for that experience versus most other data products are built for some arbitrary data set being plugged. into it because the nature of data sets has been very heterogeneous.

So in general, we're just trying to make it easier to do more experiments, do more building and learning, and in various different ways. So yeah, that's what we're up to. Super exciting. So where do people go to kind of learn more about June or get help building their first queries? Is the Discord where everyone is at? Yeah, there is dune.com where you can see all the dashboards, see what's trending, all that good stuff. We have a community tab on top right next to the new query button where you can, you know, see all the learning material, find the discord and then connect with the community. We're at June Analytics on Twitter, where you can see a lot of what's being built. We have a weekly newsletter with a lot of great breakdowns and analysis based on data, for instance, the new

unispop research post on like what happened with the uni token air drop and what that turned into on chain so yeah we're pretty accessible it's if you're curious and interested in data i think we're pretty easy to find and hopefully we'll make your life easy and exciting yeah awesome thanks so much frederick for coming on i think this is a great place to wrap up so yeah we're excited to try out the API and, yeah, hope to have you on at some point again in the future. Thank you so much for having me. And congrats on, I don't know, 478 episode or something. Is that what we're at?

That's a very impressive consistency. I think it's... 472. It's like almost 10 years now. We've missed one week. This is still a host spot for us, but yeah, we missed one week. That's so impressive. It's not to be underestimated how hard it is to do something for a long time.

So well done. Congrats to you guys. Thank you, Frederick. Thank you for joining us on this week's episode. We release new episodes every week. You can find and subscribe to the show on iTunes, Spotify, YouTube, SoundCloud, or wherever you listen to podcasts. And if you have a Google Home or Alexa device,

you can tell it to listen to the latest episode of the Epicenter podcast. Go to epicenter.tv.com. For a full list of places where you can watch and listen. And while you're there, be sure to sign up for the newsletter. so you get new episodes in your inbox as they're released. If you want to interact with us, guests or other podcast listeners, you can follow us on Twitter.

And please leave us a review on iTunes. It helps people find the show, and we're always happy to read them. So thanks so much, and we look forward to being back next week.