🚀 SUBSCRIBE TO NEWSLETTER: http://bankless.substack.com/ ✊ STARTING GUIDE BANKLESS: https://bit.ly/37Q17uI 🎙️ SUBSCRIBE TO PODCAST: http://podcast.banklesshq.com/ 👕 BUY BANKLESS TEE: https://merch.banklesshq.com/ ----- 💪BECOME A BANKLESS PREMIUM MEMBER: http://bankless.cc/membership ----- GO BANKLESS WITH THESE SPONSOR TOOLS:
⭐️ AAVE - BORROW OR LEND YOUR ASSETS https://bankless.cc/aave
🚀 GEMINI - MOST TRUSTED EXCHANGE AND ONRAMP https://bankless.cc/go-gemini
💳 MONOLITH - GET THE HOLY GRAIL OF BANKLESS VISA CARDS https://bankless.cc/monolith
📱 DHARMA - MOBILE ONRAMP DIRECTLY INTO DEFI https://bankless.cc/dharma
------ AMA with Yaniv Tal, Project Lead at Graph Protocol
The Graph (https://thegraph.com/) is an Indexing protocol that organizes data from blockchains, making decentralized data interoperable and easily accessible.
Yaniv on Twitter: https://twitter.com/yanivgraph?s=20
------ Don't stop at the video!
Subscribe to the Bankless newsletter program http://bankless.substack.com/
Visit the official Bankless website http://banklesshq.com/
Follow Bankless on Twitter https://twitter.com/BanklessHQ
Follow Ryan on Twitter https://twitter.com/ryansadams
Follow David on Twitter https://twitter.com/TrustlessState
----- Not financial or tax advice. This channel is strictly educational and is not investment advice or a solicitation to buy or sell any assets or to make any financial decisions. This video is not tax advice. Talk to your accountant. Do your own research.
Disclosure. From time-to-time we may add links in this channel to products we use. We may receive commission if you make a purchase through one of these links. We'll always disclose when this is the case
No summary for this episode yet.
Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Hey Bankless Nation, we are so excited to bring you another Community Ask Me Anything. We've got Yaniv from The Graph. He is the co-founder of The Graph Protocol. We're going to dive into The Graph and what it does. Just some ground rules for you. So you may be watching this live on YouTube. If you are, you can ask questions. David and I will get to your questions on YouTube. We'll try to prioritize them the best way that we can. If you are a Bankless Premium member, we will also prioritize your questions in the Bankless Discord.
This isn't ask me anything. So David and I have a list of questions, but we also want your participation. This is a great opportunity for us to figure out what's going on with this important piece of DeFi, Ethereum, crypto infrastructure known as the graph. Yaniv, welcome to Bankless. Welcome to the AMA. You ready for this? I'm ready. Hey guys, thanks for having me on the show. So it's awesome to have you. We've been wanting to do this for quite some time. You know, I think we should start is just sort of an explain it like I'm five on the graph. So bankless listeners, we're all familiar with DeFi.
We're obviously familiar with kind of Ethereum, Bitcoin and crypto. So we've got that base knowledge. But what is the graph? It feels like it's this behind the scenes infrastructure. What does it do? And like, have we used it without knowing? Yeah, I bet that you have. So, you know, the graph is an indexing protocol and the explain it like I'm five is that, you know, we organize all of the data from blockchains and make it easily accessible to applications. applications. So the same way that you wouldn't find a topic in a book by just kind of scanning page to page, you need to have a table of contacts or an index that organizes that data sorted so you can easily find what you're looking for.
Computers work exactly the same way when you use applications, they're querying data from some underlying place. And before the graph, people had to build custom, and proprietary centralized servers to index that data to make apps load quickly. And now there's, you know, we've launched the graph network which is a decentralized network where anyone can participate and run an indexer to help index the data for DeFi and web three. And people can now build fully decentralized applications that query that data from the graph. And it's a really just core piece of infrastructure for building decentralized applications.
Okay, super cool. So, you know, Google is famously sort of an indexer for the web, right? There's all of these different like websites that are pushed out there and they built a lot of tech over the years to help us find things on this permissionless system known as the internet. They're kind of like an information indexer, if you will. Is this kind of what the graph is doing for blockchains and for crypto? It's just like searching through it, categorizing it, and making it easy for applications to retrieve the data that they're looking for? Yeah, I love that analogy because we do the same thing, right?
Indexing, organizing all of this information. So there's some key similarities like that. And then the difference is that Google exposes that as a search box. You go to google.com and that's the interface is this kind of search engine, whereas the graph exposes all of that data through APIs called subgraphs. And APIs are how developers build applications and how they interact with other computer systems. So by having all of these open APIs built by developers for developers where all of that data is indexed, we can actually serve custom applications. So instead of it just being a search box taking you to a website, you know, you can have apps like Uniswap.
info or the Synthetics Exchange or the Decentraland Marketplace. You know, all of these are very custom applications, but they all need to get their data from somewhere. They need to allow users to search, filter, sort, you know, find relevant data, and they can do that all with a graph. So, Yaniv, you said that we might be using the graph without being totally aware. What are some examples of where the graph is actually used for the typical bankless listener, the typical DeFi user? Do you have some examples for us? Yeah. So, you know, in the DeFi world, you know, some of the projects I'm sure listeners have used are, you know, Uniswap.
info, Synthetix Exchange, Aave, Balancer, just to name a few. And, you know, I think at this point, you know, talk to basically just reach straight, you know, for the graph from day one and it helps them, you know, get their products to market just significantly faster. So this is uniswap.info. I don't know if you could see that, but I'm pulling that up on the screen for everyone who's watching over YouTube, right? And so I'm on the overview page. I'm looking at all of kind of the liquidity on Uniswap, all of these various data points. This is what you're talking about, right? So you're doing, the graph is doing the indexing and the data retrieval portion for all of these graphics, all of these dashboards.
Is that correct? That's right. So, you know, it's loading the data for the charts for like those top tokens, you know, you can sort by liquidity or by volume, right, by clicking on the column headers. So that's, you know, a query with, you know, sorting, you know, if you click in on the tokens or on the pairs, you know, there's kind of, you know, historical transaction data and volumes and all of that data is coming from the graph. Okay, very cool. And the reason this is important, One reason this is important is because, of course, open finance is better, as we say, than the traditional financial system, because the data is accessible to anyone on chain, basically.
So if you want to see who owns a specific token, which Ethereum addresses do that is, or if you want to see how much value is collateralized in the DAI that you're holding that underlies a collateralized debt position. Like this data is available on chain, but it's not necessarily accessible in a user interface that you could see. And the graph is essentially making open finance accessible in user interfaces in places where people can see and they don't have to kind of go digging for it and run their own queries and find it, right? That's right. And we think it's important that all of the infrastructure for DeFi and Web3 is decentralized.
We want to move away from these situations where a server goes down and an app stops working or people take dependencies on specific kind of centralized points of failure. You know, anytime you have like one group of people decides, you know, these are the top DeFi protocols or these are, you know, the metrics that you should be looking at to evaluate, you know, a synthetics or an Aave. You know, those are decisions that really should be transparent and that if people want to like, you know, fork them, they can fork them. And that, you know, anyone can build applications easily on top of this data.
And, you know, information works in these kinds of layers where, you know, it's constantly aggregated and repackaged. And so the more that these layers are, you know, stable infrastructure that teams can build upon, you know, the more amazing things we can build. Okay, so now we get to kind of this question. So people have pointed to application services like Infura as being a centralization vector for Ethereum and for DeFi, of course. Is the graph similar? Like, how does the graph prevent becoming a centralization vector in this decentralized open finance thing that we're building? Yeah, so that's exactly why we built the graph network.
Initially, we launched the graph as a hosted service, and that was so that developers could start building on the graph. And we wanted to make sure that we were building something that was useful for developers. But we were aware of this risk from the very beginning. And it's why from the beginning, we set out to build the graph as a decentralized protocol. And last December, we actually launched the decentralized network. So in this network, it's an open permissionless network. Anybody can run an indexer to provide the indexing and query processing service to the network. As a client, you can choose who you want to connect to.
And so all of that choice is on the client side. And that specifically makes it so that there is no single point of failure, single kind of like, you know, choke point, decision point on, you know, what is valid data, how should the data be indexed. And ultimately when you bring this up to the API layer, you know, that's kind of the top of the stack. That's where you're asking the question like, you know, what NFTs are for sale, sorted by, you know, price. And you don't want someone to be able to just, you know, change the results, sensor, you know, make these types of decisions on their own.
And so you want to have some guarantees that the data that you're getting back when you're running that type of a query is correct. And you just couldn't do that without something like the graph with this indexing layer. Initially, when Ethereum first launched, there was kind of this vision of light clients really being the main way that you would interact with everything and light clients are really great because you can basically verify proofs on the clients you know, proofs on the client that some value that you requested is correct. But you can only do that with a very narrow subset of kind of operations.
So, you know, about an Ethereum balance, you can verify that, but you can't do that for kind of like arbitrary views on data, which all, you know, applications produce custom arbitrary views on data. And so if we're going to have rich applications that do all of the kinds of things that we're used to traditional web or mobile apps doing, we need to be able to actually transform and consume much broader sets of data that have been transformed and organized. And so you can do that and still have it be verifiable on the graph. You know, I think it would be helpful if we talked about what would what are the potential attack vectors if the services that are provided by the graph are remain services provided by centralized intermediaries?
Like what what are the attack vectors? What are the risks if centralized services are become the dominant indexer for DeFi? Like, are there like attack vectors that could be exploited or where what could go wrong if this service that is provided by the graph remains in the hands of centralized intermediaries? Yeah, it could be everything from raising prices and suddenly everything starts to get expensive, making censorship decisions around either this DAP just isn't allowed or even more granularly than that. Imagine you have an NFT marketplace and then there's a certain set of NFTs that basically the service provider just says, like, hey, you can't have these NFTs.
And they start making those types of decisions, you know, kicking off developers. And then, you know, just, you know, so these are all types of like attack vectors. But ultimately, I think the problem with centralization is that these things require a lot of development and a lot of effort. We're at the beginning of building out an entire new financial and information infrastructure. And, you know, developers don't want to contribute to something that is owned and controlled by a corporation that can make these types of decisions, just like on their own, like we've, we've done that, you know, we've been there and, and, and now we're going to rebuild the stack so that it doesn't have the same pitfalls and problems.
And so by having a decentralized protocol, what you have is a shelling point where people know it's worth investing my time, you know, to this thing because I know it's going to be around in 100 years. You know, it's not going to be co-opted. And that allows just a tremendous amount of development to get concentrated for the public good. You know, something that anybody can then use because, you know, everybody can contribute to it. Maybe let's talk about the components that network participants contribute to the graph and how the graph uses these components to offer its product. So as an individual, what can I offer to the graph that the graph needs in order to be able to create its product?
Yeah, it really depends on your skills. I mean, the main roles in the network are indexers, curators, and delegators. The indexers are the ones that are running the nodes. So if you've like run an Ethereum node, for example, and you know how to do that, then you could look at becoming an indexer on the network. We have curators. Those are the ones that understand the data that's being indexed and are organizing the data on the graph. They could be subgraph developers, for example. So, you know, anyone can build a subgraph that indexes data and makes it available over GraphQL. well. Or the curators can also signal on subgraphs to basically evaluate the subgraphs, figure out which ones are highest quality and should be used by applications.
And then they can signal on those subgraphs to make it easier for others to find it. And it also impacts the economics for the indexers. And then finally, we've got the delegators who help secure the network, but they may not be as technical themselves. So they can just find indexers that they think are doing a good job and just delegate their tokens to those indexers, which allows the indexers to earn more fees and rewards, which they pass back to their delegators. Yeah, let's talk about the token and the economics of the protocol. So every single protocol that has a token is using this token to manage incentives.
How does the graph use its token to manage incentives and perpetuate the network? Yeah, so it's used for staking by indexers, delegating by delegators, and signaling by curators, and also as query fees by consumers that are querying the network. So, you know, first and foremost, it's a staking token. It creates economic security. So the indexers that provide the service in the network have to stake tokens so that if they misbehave, they can be slashed. They then set their query fees also in GRT and indexers can choose what subgraphs they want to index and they can set their prices very granularly.
And those fees are paid using state channels. And, you know, we're, you know, I believe the largest, like, large-scale deployment of state channels in production on Ethereum. Wait, I had no idea. Why state channels? Yeah, that's pretty cool. I guess that makes it more decentralized. Yeah, tell us about that. Yeah, so we did a partnership with the state channels team in ConsenSys R&D and Connects. And, you know, we evaluated the different options of doing payments, you know, in the graph, you, you have this very like many to many style relationships of payments where you can have like lots of different consumers and lots of different indexers, and the actual payment amounts are quite small, you know, just a small fraction of a cent per query.
And we wanted to ensure that, you know, the queries themselves can differ significantly in expense, you know, most queries are a very small fraction of the set but you could have like a larger like analytics style query that, you know, could take a second to execute or something. As a service provider, you want to know that you're going to get paid, you know, if you do that work. And so we have a construction called, you know, conditional micro payments. we also use this term like way blocks. And it basically ensures that when you get a query response back from an indexer, it comes back with an attestation.
And if the query response that you got back is incorrect, then you can submit a dispute on chain and the indexer is liable and can be slashed then if they gave you the incorrect response back. So that whole kind of interaction happens inside of a state channel. It's a state channel game. And it allows you to, an indexer can be doing millions or even billions of queries a day. It's very fast and cheap. And then they just settle those channels on chain, whatever frequency they want. It could be daily or weekly and collect their funds. This is amazing. So I had no idea that you guys were doing this with state channels.
So this is really making it like scaling a decentralized and like payment layer for you that's really necessary for your work token, your entire token economics, and you're keeping it decentralized using state channels. So how many transactions per second can this support? Do you need additional scaling tech to support this payments use case or state channels give you what you need? State channels give us what we need. So we could scale up to billions or more queries per day on this tech, and it works great. I think we are evaluating layer two scaling solutions for some of our core protocol smart contracts because interacting with the protocol itself, you're subject to all the same kind of gas issues as other protocols.
So we'd like to bring those costs down for when you do have to make on-chain interactions. But for the query processing and the query fees, the state channels is actually working really great. That's probably a combination of both the need for micropayments as well as the need for immediate payments. Is the immediacy of payments a nice feature that state channels offer or is that something generally unrelated? Well, you do want the property that once a query has been served, basically the transaction is finished and the parties know that the value has been transferred. Right. So waiting for an Ethereum block is too slow.
Right. Yeah, exactly. I mean, if you think about most applications, if you're using Facebook or Twitter, you're consuming, you're reading so much more data than you're writing. You know, like the number of tweets versus, you know, you scrolling a feed. So, you know, the querying is happening like constantly all of the time. And so that just needs to be fast and instant. So this token seems to be, in my mind, it maps on to, I think, the Chainlink token model. And so correct me if I'm wrong, but like there is people that are staking on a promise to provide accurate data to the network, there's people that are consuming that data that are paying fees in the native token.
Um, and, and then there's also the, the, uh, the, the protocol itself, which is, uh, an economic, a crypto economic protocol that's meant to provide data to the world. Um, is that a fair comparison? Would you say? Yeah, I think it is, you know, I'm excited for the chain link staking to go live. Um, cause that's, that's not in production yet, but I believe it's a similar sort of model. And I love the Chainlink also comparison where they help bring real world data on chain and we help bring on chain data to the real world. So we're very much just kind of like dual middleware protocols. Yeah. Do you guys have a green frog army though?
You have a green frog army or no? No, we don't. We're currently recruiting. So we need you. What? So speaking of that, right, it's like David was getting to this earlier, and I think you touched on it. We actually did a bankless tactic on this about like how to work essentially for the graph network. So what you just described is basically the graph is paying for what the graph network needs using its token economics, right? And the things it needs are curation, delegation, but also indexing, right? And so like a number of folks that are active in these DeFi protocols, they'll have done some work on curation, right?
It's fairly simple, not super technically complex. And then also delegation is something we do in almost every protocol. That's staking, right? Can we talk about this indexer function, though? This is more technical, but this to me is the question of decentralization. Because if it's really hard to run an indexer, if I have to be like a super sophisticated DevOps with my own data center in my backyard, right? That's less decentralized than something like ETH2, which is I can run a validator on my laptop hardware, right? And I'm guessing it's not like we can run an indexer on our laptop hardware necessarily.
It's maybe somewhere in between. But can you talk about the complexity of actually running an indexer? Do I have to be a sophisticated network operator with my own data center, all this provisioned? Or is it a bit more do it yourself? And how does that scale? Because that to me is a centralization vector in the network. Yeah, so, you know, I think of decentralization for a network like ours differently from, you know, as I would for like a layer one blockchain. You know, I think Ethereum made a great set of trade-offs to, you know, be maximally decentralized for validation. And I think for a layer one chain, you know, that is really important.
For what we're doing, it's kind of like a service protocol. That's what we've coined it, different from like a layer one where, you know, we're providing a service to the network. It's indexing and it's very kind of heterogeneous. So each indexer chooses which subgraphs they want to index. They're not all doing the same work. And then the client chooses which service provider they want to do business with. So they have complete control. I can say, you know, I really don't trust these indexers over here. I just don't, you know, for whatever reason, I want my money to go over here. And you can choose which service providers you do business with.
And so because the user's in control, you know, they, you know, it is kind of, you know, produces choice from that perspective. And then, you know, for networks like the graph, what's really important is kind of like marketplace dynamics. You know, you want it to be easy for new service providers to provide their service. You want there to be lots of choice, consumer choice. So consumers can choose which indexers they want to do business with. And you want to have competitive prices and a very high quality of service. So for that, it makes a big difference whether an indexer can serve a query in 0.
1 seconds versus two seconds. So you actually do want this level of kind of competitiveness and, you know, serving data and processing data is something where you want folks to have BP setups. You want it to be, you know, so that they can provide that service really efficiently. So generally, the people running indexers are ones that have access to faster servers. We want it to be something that anybody can do so that barrier to entry is fairly low. If you have a server set up, a reasonably fast computer in your home, you could actually be an indexer. But we also want there to be ways for people to compete on the higher end of the spectrum where they're providing a faster service and maybe it just costs a little bit more.
I think one of the most elegant things about Ethereum and what good applications can do on Ethereum is let the free market come into their protocol and leverage market-based incentives to produce a quality product. And I'm getting vibes of like sharding here. So Ethereum scales by sharding out computational resources to 64 different shards. And so we get 64 times the scale because each person doesn't have to process all of the data, only a subset of the data. And that seems to be a similar pattern that the graph is going here where and they're actually you guys are actually making this even more free where you can just say, like, you know, process whatever data that you want, but you'll only be compensated if the market actually wants that data.
And in addition, I would imagine that there is certain amounts of data about Ethereum that's just simply more in demand than others. Perhaps like data around Uniswap might be more in demand than data around CryptoKitties. And it seems to be like the token would be the substrate that allows that free market to become expressed. Maybe you can talk about how data, in-demand data, will become better served over time and allow kind of the free market to come out and produce a quality product here. Talk about a little bit of the free market relationship with data that's in demand. Yeah, those are great insights.
So, you know, right now, subgraphs are kind of the units of composition. So, you know, there are, you know, there's a Uniswap subgraph, synthetic subgraph. And, you know, soon people will actually be able to, like, compose those subgraphs into, like, aggregated subgraphs. So you have this unit, which is the subgraph. From there, an indexer can choose which subgraphs they want to index based on, you know, like you said, demand, what the current prices are in the market. it, you know, if they think that they can do it, you know, more effectively than others. So they choose the sub graphs and then they can set the prices granularly.
So they can see like these types of queries are really expensive for my database to run. I'm going to run, you know, charge a little bit more, or, you know, I've been able to cash this data. I know like every time someone loads unisoft.info, they make this type of query. So I'm just going to like cash it and I'm going to serve it super fast, super cheap. So the indexers can set their prices really granularly in this query market. And I think that kind of a flexibility is really exciting for this type of a thing because databases are notoriously difficult to optimize. Most web companies, they'll have dedicated DBAs or database administrators whose job is to make sure that an app developer doesn't accidentally run a query that takes the whole database down.
And it's actually really easy to do that because for certain applications, the types of queries that you run, some can be really simple and take no time at all, and then some can take a minute and bring the entire database down. So to have actually this kind of open market for indexing data basically allows you to have this distributed decentralized DBA team sitting there optimizing database access, figuring out how are people accessing data today, what types of applications are they running, they can go in, make the necessary optimizations and make sure that basically all of the dApps on DeFi and Web3 are super fast, performant and cheap and the developers just don't need to think about it.
How does this help solve the Ethereum ecosystem's dependency on Infura? Is this related? Yeah, well, you know, the graph is kind of a, you know, it's a layer on top of, you know, the blockchain node providers themselves. So, you know, an indexer could choose to use Infura, you know, as its Ethereum node, or they could run their own. We have noticed that most of our indexers, you know, chosen to run their own archive nodes, which I think is great for decentralization. So they're actually validating the chain themselves. And then, you know, running this indexer on top. So, you know, ultimately, I think, you know, we can move to a world where, you know, there are, you know, hundreds or thousands, many thousands of indexers.
Each one is validating the underlying networks whose data they're processing. They're serving it up over the graph. And then the clients can, in an automated fashion, do the indexer selection. And I think you end up with a much more robust and decentralized infrastructure layer. So Yaniv, you started The Graph, like co-founded it a couple years ago, is that correct? Like 2018? Yeah, we started on it kind of Christmas break 2017. Okay. All right. So you obviously had an inkling that crypto blockchain would be important and that there would need to be a lot of data that required indexing. But did you see it playing out in this way?
Like, is all of this surprising to you? Or did you anticipate it going exactly the way it's gone? Yeah, honestly, it's kind of scary how little the vision's changed. And by little, I mean, basically, yeah, we, you know, the vision that we saw is nothing's changed about it. I think we didn't expect the big crash. We didn't expect some elements of it to kind of take as long as it has to kind of progress. And if you asked me in 2017 how long I thought it would take to build the graph, I probably would have told you like a year or something. But it was a lot of work. I think that's the thing that I've really appreciated is just how much work goes into building a protocol, how much work goes into getting developer adoption and building tools that developers are excited and happy to use.
But the vision is played out basically exactly as we initially set out for. So how about the applications, right? So in a way, the graph is kind of privileged because you don't really have to take a bet on what applications, even like what areas on top of something like Ethereum become successful, right? Is it like in the end of 2017, everyone was talking about Web3, right? Decentralized, you know, Web2 type companies, right? You know, it turned out that the first major use case on Ethereum was more money type use cases, more like DeFi type use cases, at least that's first. But you guys didn't really have to take a bet on which applications, which use cases would be successful on top of this infrastructure, because you're like the, you know, picks and shovels type of infrastructure type of play.
But let me ask you, were you personally surprised about DeFi and how it's gone or the use cases that have popped up in the crypto sphere? Or did that play out how you thought it would? Yeah, yeah, that, that was surprising. You know, I think, you know, I remember when there was like a billion locked in DeFi. And that just seemed like enormous. And already we see like, you know, how quickly we've grown from there. So, you know, that came by surprise, like a storm and, you know, really excited to see how well DeFi has progressed. I do think we thought that Web3 would, you know, happen sooner. And when we were getting started building, it's like we were focused on our protocol.
And we just assumed that all of the other protocols were doing the same and that they would just come together, you know, relatively quickly. And I think, you know, by the end of 2018, it was clear that like a lot of these other things outside of our control were just going to take longer to kind of get, you know, get to market. But I think, you know, you know, people kind of, you know, they overestimate what they're going to see in the short term and they underestimate what they're going to see in the long term and i think that actually a lot of the initial ideas that people had i mean even things like state channels like we were talking about like in 2017 people were talking about these things like they were going to happen immediately you know plasma is about to launch and all of these things are are about to launch they've taken a little bit longer but you know uh it really feels to me like 2021 is the year you know for for web 3 and that there through all of these incredible protocols that have really been under development now for like three years and are just kind of getting to market.
And I think we'll be able to put these web through protocols together and actually build, you know, a comprehensive, you know, stack that allows you to build just incredible applications, but where all of the data is verifiable, right? Where it's actually on this, you know, more trustworthy foundation that, you know, we can all build on top of. And I think that the timing is right. I mean, you see all of the stuff happening with censorship and centralized control, from governments to large corporations getting pressured. Society, I think, has woken up to those risks. And at the same time, the financial system, it's so funny watching what's happening with like GameStop, where you see these hedge funds literally just distorting the system and corrupting it and using it to their advantage.
And I feel like ordinary people have just gotten the brunt end of the stick for so long, basically just being told, you know, you got to play by the rules and like, you know, this is the game. And then you just see, you know, Wall Street just taking advantage. And so all of these headwinds are kind of colliding. It's a perfect storm right now for DeFi and Web3. And we just need to put these last few pieces of the stack together so that people can really build great decentralized applications. And that it's simple, you know exactly like, this is what it means to be a decentralized application.
This is how we do it. This is how it plugs in. And once that foundation is in place, I think we're gonna see just like an explosion in the number of dApps because I think the world's ready for it. So DeFi hit product market fit, but I've read it in some of your previous writings. It sounds like you are still bullish longer term on this Web3 decentralized kind of internet, decentralized social type movement. Is that the case? Absolutely. I think it's going to come sooner than people realize. And, you know, I think this is the year that we basically put the Web3 stack together. And then I think, you know, we're going to be ready.
It's going to happen. Yaniv, I got a question from YouTube that I want to get to as well as a few others. The first one is, is the graph married to Ethereum or are there other L1 blockchains that the graph is working with? You know, we do think it's going to be a multi-chain future. You know, I think Ethereum is going to be, I would imagine, you know, the most valuable kind of like primary chain. But I think there are different points in the trade-off space. And just even for resilience and decentralization sake itself, it's good to kind of have backups and, you know, different alternatives. And so I think ultimately, people can choose where they deploy what contracts based on their needs and that we can stitch all of these things together into a larger decentralized platform.
How about Layer 2? How does that kind of rock your world? So we've got Synthetix and they're just testing out Layer 2 on Optimistic Rollup. up. It looks like we've got things like loop ring as well. It looks like Uniswap may do that. So in the future, I would imagine an index network like the graph will have to pull from not just Ethereum main chain, but also all of these other layer twos. Does that really rock your world or do you have the infrastructure set up for that and are ready for that challenge? Yeah, we're ready for it. So, you know, that's, it's definitely coming up and we're super excited about layer two i think it's gonna allow just so many more different types of gaps to get built with ethereum's security um and and we're super excited for that so yeah i mean the graph is gonna have to index data from you know eth1 from e2 all the different shards all the different layer twos all the different you know side chains and other chains um uh storage networks you know like Filecoin, Arweave, SIA.
So it's just a tremendous amount of data that already exists in the Web3 ecosystem and there's going to be a lot more on the way. Here's another question from YouTube about the token distribution. Can you tell us about how the token was distributed to the world? Yeah. So we did a number of programs to help kind of uh uh you know bring out the the the graphs community and and to recognize people for the work that they contributed to the network um so we did a test net um to get the initial set of indexers on board we had 200 indexers from all around the world um they got about uh you know three percent of the network um we did a curator program um that was uh you know the uh either subgraph developers or people that wanted to like learn how to you know curate and organize data on the graph and that was just an incredible experience um you know i think it was something that was really brand new and that i hope that other protocols uh key off of you know is basically this you know massively multiplayer online vocational training program by teaching people you know new skills you know to perform a new job you know how to work for a decentralized protocol and you know i think that is just the future right just you know hundreds and thousands of permissionless jobs doesn't matter where you are in the world you know here's the the educational materials you know here's how you know the skills that you can um learn to perform this job and then have at it nothing standing in your way.
You don't have to ask for permission. So we did that with curators and we allocated an initial distribution of tokens to them. We did a token sale. How big was that distribution, if I can ask? What was the percentage of the, what were you just talking about, the distribution for the vocational tuition? The curator program was about three, yeah, the curator program was about 3%. percent uh and then we did uh the uh um the token sale uh the the public token sale portion was uh was four percent and that one we also did very differently we did a custom uh token sale design built our own smart contracts our own ui it ran on the graph which by the way it ran flawlessly and you know a lot of these token sales that are done on centralized platforms you know there have been a lot of hiccups where like you know the sites just crash and there's all of these questions about like who got in and why and and so actually really showing that like the d5 and web3 infrastructure is just already better than centralized you know platforms for a lot of things um so in that sale uh the way that we designed it was pretty different where we had these three phases and we you know we kyc'd everybody you know because we had to and uh there was a question what excites you about DeFi and Web3?
How would you describe the value of the graph? And based on the answers to those questions, we prioritize folks that really understand what the graph is and we're aligned with the mission and wanted to contribute. And we prioritize them for the first phase where they had a guaranteed allocation. It was small allocations of a thousand to $5,000, but that way we were able to get 45,000 individuals in. It was kind of this anti-civil and these are people that are all really like aligned with the mission of the graph. So I felt that that design works really well for distribution. How did you go through 45,000 people?
Who had to do that job? Oh, sorry, 4,500. Oh, okay. Okay. Yeah, yeah, yeah, yeah, yeah. 4,500. were 20,000 applications. So not everybody was able to get in. The second phase sold out in 10 minutes. But yeah, all the transactions were from the first few seconds. Awesome. That's a pretty cool. Go ahead. Go ahead. Yeah. Just so you know, it was very hands on and it was a lot of work from our business team to kind of help coordinate. But really, our general approach, which I think was kind of unique, is, you know, we wanted to reward real contributions, right? People that understood, you know, what the protocol is trying to accomplish, and we're actually contributing in some way.
And there's lots of different ways to contribute to a protocol. But, you know, that kind of a system where we're basically, the initial team is playing a pretty active role in the beginning, you know, using maybe some amounts of judgment, even, you know, which may be a subjective, but I actually think that, you know, some amount of judgment in the early days is, I think, a lot better than kind of like blanket, you know, systematic lack of judgment. And I think it does allow the community members that are really active to be recognized, and rewarded. And I think it's been a big part in allowing us to actually build like a really, you know, kind of diehard community.
I totally agree. I think the surgical use of token distribution, especially at Genesis, if any other time, specifically during Genesis, the surgical use of token distribution, I think just makes the most sense because it's that's the first impression of the protocol, right? You need to incent the right people at the very beginning, most of all. And so I think top-down choice around the token distribution at Genesis, I think makes a ton of sense to bootstrapping the right community. What about the remaining supply of tokens? Are there like VCs, VC investors, or there are their team allocations?
Like where are the bulk of the tokens laying? Yeah. So I believe it's somewhere around a third was for the initial kind of SAP supporters. So, you know, we took in a first capital in 2018 and it was, you know, two to three years of development to even, you know, bring the protocol to market. So that was funded through pre-sales of the token. The initial team members have an allocation. Edge and Node has an allocation. So that's now the initial team is, has this new entity Edge and Node. And we've got some announcements we're going to be making about our participation in the ecosystem. So stay tuned for that.
The graph foundation has the largest kind of, you know, remaining portion there. It's about 20% is for the foundation to be distributed via grants and hopefully eventually a DAO. So the grant applications, I believe for the first wave just closed, but we got just a ton of really great quality submissions for folks that wanna contribute to the network, either development or community building, lots of different ways for people to contribute. So the graph foundation, which is led up by Eva Balin is going through those submissions And we're going to start distributing grants. And those are the big buckets.
Fantastic. Let's talk about the future roadmap for the graph, what you guys have on the horizon. And by the way, people watching on the YouTube, A, thanks for being here. B, last chance to get your questions in for Yaniv before we sign off. Yaniv, before those questions come in, maybe you can talk to us about the future roadmap of the graph. Where is the next 6, 12, 18 months of development taking us? and what are you guys trying to roll out in that time? Yeah, so, you know, first thing on our mind is, you know, the full transition over to the decentralized network. You know, as of the end of last year, we were processing 10 billion queries per month on our hosted service.
And and now we're kind of transitioning, you know, over to the decentralized network. So we're going to be launching a graph gateway, which makes it really easy for developers to publish their subgraphs to the decentralized network and pay query fees in GRT for those subgraphs. And there's also going to be a new graph explorer for people to browse and discover, you know, the main net subgraphs. That's going to be what creates the full kind of self-service experience for developers to really get running on the network. And so there's going to be this kind of migration process from the hosted service to the decentralized network.
We're also going to be expanding multi-chain. That's going to be a community effort, working with a lot of people, you know, in the community to actually bring that to fruition. So excited for that kind of expansion. And then generally, you know, in 2021, we really want to be focused on, you know, building out that Web3 stack and really getting it to a point where there's a really cohesive platform that's super easy for developers to build on. And that gives you all of the properties that we want from decentralization. So that's that's going to be the focus for the rest of the year. Can we talk about this?
Because I think you've got a good take on this, you need to, is what about the roadmap for this whole thing that we're calling open finance and or Web3? So it seems like we have some of the building blocks here. So Bitcoin started as sort of a store of value. We've got that thing, that primitive. We've got Ethereum. We've got Ether, which we'd argue is also sort of a store of value. And all of these money Legos on top of Ethereum. So it feels like we have this DeFi set of primitives. We almost have like decentralized banking and money for the space. Then we've got the graph, which is a decentralized indexer.
Then we have things like IPFS and Filecoin, which are doing decentralized storage. What other pieces do we need? If you're looking at the high level roadmap for this whole decentralized stack, what else is missing? Yeah, all of the pieces are being worked on. And there's a few questions about kind of how they integrate together. Once maybe that weren't in that list that I would add our identity and reputation, I think are really key primitives. You know, obviously things like layer two scalability, but there's I think the Web3 stack is going to be about, you know, 20 protocols. some that are doing the same kinds of things and you can choose, you know, do I want to use SIA or Filecoin or, you know, for storage or, you know, there's some duplication there, but ultimately I think, you know, it's going to be around a dozen to 20 protocols, you know, some, a few different like kind of components and then they kind of need to be stitched together.
And I think, you know, we're basically still at the phase right now where we're just kind of assembling these Lego blocks. I think we also need like design patterns. So like staking and delegation, for example, I think is a fairly mature as a design pattern. You know, you know, the way that we use it in the graph, for example, like it works really well. And there are enough protocols that do staking where we kind of like understand the design space. You know, governance, you know, is, you know, the like the compound governance contracts have emerged as a bit of a standard. but there's still different kind of templates and design patterns that we need to come up with that makes it really easy to build new protocols.
And ultimately I think that we'll get to a point where, to build a protocol, you'll ask yourself a few sets of questions and you'll kind of pull from the right kind of building blocks to kind of like assemble, here's how the protocol is generally going to be orchestrated. There might be things that you need to customize, But the design patterns, those Lego blocks will be well established. And once we get to that point, I think we can go from having, you know, maybe like a dozen or two dozen, like well thought out protocols to having like tens of thousands of protocols. And ultimately I think that's going to be the next wave that we enter where we actually start to decentralize everything.
You know, we start using these tools to build protocols for every part of the economy. All of that is gonna be organized on the graph. People will be able to build decentralized applications on top of that. But we go from basically being in the lab, tinkering around with toys, to actually just rewiring how society is organized. And I think it's gonna come out as fast. That's pretty cool. I got one final question from the YouTube I wanna ask. Are there any other graph competitors out there? Who else is trying to serve the same sort of purpose? Well, the main kind of thing that we compete against is teams building their own centralized indexing.
And I think because you kind of couldn't really build fully decentralized applications over the last several years, it's kind of become acceptable that teams say like, know what we're going to have to run a server anyways let's have the server do more and more stuff and you know it it it's it's challenging because in the short term maybe nobody's calling them on it but but they're kind of building themselves into a corner because ultimately we are going to move to a world where like you just don't want teams to run servers and it's going to be kind of like a you know a red flag if some application that you're using requires like a team to run a custom server.
And so the sooner we get to a point where teams are building their protocols and their dApps on purely on open protocols, you know, the sooner we can, you know, get to the future. Yeah. And Yaniv, do they have, do you have any metrics for us about like how used the graph is like queries per day or queries per minute or something? Is that a, what's the right metric to evaluate how used the graph is being? Yeah, the latest metric that we had was 10 billion queries a month. That was on our hosting service, which we're in the process now of transitioning over to the decentralized network. Also, number of subgraphs is a big one that we track.
I believe that the latest is 7,000 subgraphs deployed to the hosted service. We have hidden subgraphs that don't show up in the Explorer. And we count that in that figure. So 7,000 subgraphs published. And we also track kind of number of developers, which are in the many thousands. Okay. So with queries per month and number of subgraphs, where are your end of 2021 goals? Oh, yeah. I actually don't have a goal for this for 2021 because I think that the nature of the transition to the decentralized network is going to be potentially a bit of a discontinuity. So really what I'm kind of focused on, whenever we do something new at the beginning, we're focused on much more qualitative sorts of measures.
So like, you know, is the query market developing in a healthy way? Are the indexers happy with the way they're able to set their prices? Are the consumers happy with how the prices are working out? You know, how can we lower the friction of kind of migrating over and just like educating people on what it means to actually be running on this like decentralized infrastructure. So, you know, for this year, I think we're going to be focused on a lot more of this kind of like qualitative sorts of metrics. 2021 is the year of decentralization for the graph. Yeniv, thanks so much for joining us on this Bankless AMA.
It's been a pleasure. I know the community benefited a tremendous amount from your insight and thanks for building what you have built in this protocol. Thanks guys. Really enjoyed the conversation. Guys, if we were doing a regular podcast, I think we might want to subtitle this, your kids will be working for protocols because there are so many different work opportunities out there. David and I, we talk about all the time that we actually work for protocols ourselves. And there are opportunities to get plugged into the graph. We will include a link to our tactic on how to become a curator, delegator, and indexer on the graph in the show notes of this video, in this podcast.
Guys, this has been a Bankless Community Ask Me Anything. Of course, none of this has been financial advice. There are risks out there. You could lose what you put in. DeFi is the frontier. But thanks for joining us. This has been an episode of Bankless. And if you are still watching on the YouTube, thanks for being here. Make sure to like and subscribe the video and stay tuned to the Bankless channel for future AMAs like this one.