GRTiQ Podcast: 47 Data, Web3, and The Graph

The following podcast is for informational purposes only. The contents of this podcast do not constitute tax, legal, or investment advice. Take responsibility for your own decisions, consult with the proper professionals, and do your own research.

Ricky Eslapon (00:21):

The very last thing I’ll say on this topic is that I’m also a really big believer in people being able to do what they love. I truly believe that you’ll be able to create more value if you’re able to follow your passions. I think web3 is also a clear enabler of that.

Nick (00:34):

Welcome to the GRTiQ Podcast. Today, I’m speaking with Craig Tutterow and Ricky Esclapon, both of whom are data scientists at Edge & Node, a core dev team working at The Graph. If you’ve seen some of the recent analytics and data about activity in The Graph, then you’ve seen Craig and Ricky’s work. Both Craig and Ricky offer an incredible view into the topic of data, both in The Graph and in web3.

(01:31):

They also share important ideas about how web3 is changing the data landscape, why it’s important, how The Graph is leading the way. As you’ll hear, Craig and Ricky both have divergent backgrounds and entry points into web3, showing, once again, the type of talent and intellect web3 is attracting. As always, I started the discussion by asking Craig and Ricky about their background, starting with Craig.

Craig Tutterow (01:57):

Yeah, so I did a PhD at University of Chicago Booth, so it was the same program that Hiroki was in, who was on a few weeks ago, except I was more on the macro organizational behavior side. My affiliation was more with the sociology department. I did a lot of work on network science, so looking at in particular the structure of mediated markets and how those shape competitive dynamics and producer outcomes in specific industries.

(02:32):

After that, I went on to Google to do a postdoc in their people analytics group. I was doing network science and data science projects there. Following that, I went on to LinkedIn, where I worked in the Network Growth Data Science Group. There, I was focused on the automation of measuring network effects, so augmenting what we do in AB testing, and looking not only at the impact of a feature on the person who sees the feature, but indirect impacts downstream in terms of how their behavior affects other people’s behavior on the site.

(03:10):

During that, I got really interested in software engineering, and that was the part of the job that I was enjoying the most. I went on to join a group internally called DSPX, Data Science Productivity X, and we were working to build tools that created multifactor improvements in data science productivity. We were trying to automate some of the repetitive analyses and models that people were producing.

(03:35):

Also, give people the ability to contribute to internal corporate web applications without having to know everything about DevOps or web development. They could just use Python and SQL, the things that they knew, and contribute new visualizations, or models, or analytical output, and plug that into different core web applications that we had internally.

Nick (03:59):

How about you, Ricky? What can you tell me about your background?

Ricky Eslapon (04:01):

Yeah, so I got an undergrad in information management, which was basically business analytics work. My sophomore year, I took a course in business analytics, which really drew me towards the working with data type field. I really fell in love with working with data through that course, had an unbelievably great professor that I still collaborate with to this day. That was a really great experience for me. I just kind of started doing all sorts of data-related tasks, and I basically got an internship working for a pretty big company. I was doing a lot of manual data collection type work.

(04:37):

I started diving more into work around web scraping and data collection, data cleanup. I was also really interested in crypto, so I kind of tied that into my personal work. I started going down this path of trying to create models to predict cryptocurrency prices, which eventually turned into me creating this open source tool to, or not really open source tool, more like publicly accessible website where people could learn data science for free for the website on live data sets, which I thought was a pretty cool approach.

(05:12):

Yeah, during my day job, I worked a lot on data preparation, reporting, automation tasks. A lot of what I did actually that led me to The Graph really had to do with my own personal projects.

Nick (05:24):

Well, Ricky and Craig, part of the reason we’re talking today is you’re both working on data and analytics at Edge & Node, and of course, as listeners know, The Graph really powers much of the data and analytics currently of web3, and hopefully well into the future. What I love about both of you is the different entry points you had into web3 and into crypto and The Graph.

(05:47):

Starting with you, Ricky, you took more of a community crypto-based interest. What can you tell us about how you first became interested in crypto, and why you decided to pursue a career in it?

Ricky Eslapon (05:58):

The very beginning of my journey in crypto, I would say is, I bought some Bitcoin back in 2014, but kind of skipping ahead, I think I really got captured by everything that happened around the Ethereum ecosystem around 2017, which is pretty common among guests of the podcast, it sounds like. Yeah, the idea of autonomous organizations was really fascinating to me because I felt like web3 could solve a lot of things I didn’t really like around traditional business.

(06:22):

At the very beginning, it was definitely from an investment perspective, and I would be lying if I didn’t admit to that. The main thing that made me really passionate about the field is the potential it has for fixing things that, again, I don’t really like about traditional business. I think one big piece of that for me is the incentives that an organization has.

(06:43):

To me, building a business is about bringing a lot of value to society. Traditional business has this notion that the very best way of doing that is by only focusing on shareholder benefit, so basically, only focusing on making the investors as much money as possible, but I actually really disagree with that. I think today, we have a ton of evidence that shows us those theories aren’t always valid.

(07:05):

The way I see it, basically every company doing anything good for society is only doing it out of what we call inline self-interest, which essentially means you’re only benefiting the community because you believe the positive reputation that comes with those actions will provide a larger amount of profit to the investors than not taking those actions in the long run.

(07:26):

This idea is usually seen as a really positive thing in business, but if the only objective is to create wealth for the investors, this idea really easily breaks down when a corporation grows large enough, and the legally correct thing to do becomes to extract as much value from customers and the communities a business should actually be helping. I think one clear cut example of that is this idea of planned obsolescence, which is something that I’ve been really interested in.

(07:53):

Basically, a company is a lot better off if they can keep selling customers the same product multiple times rather than just the one time. Let’s say you’re selling a light bulb. You don’t want to create a light bulb that lasts forever, because it won’t be as profitable as having those customers come back to eventually buy a new light bulb. We see this all the time in products like iPhones, cars, video games, and the list really goes on and on.

(08:17):

Companies just want to sell you something new every year because it’s just more profitable to do that. I think this concept is particularly damaging when it comes to software and data. The way I see it is we could build really amazing things if we treat the different components as building blocks that you can just keep building upon. If we’re controlled by a traditional company, there’s always going to be the incentive, and almost even responsibility, I would say, to capitalize on the large user base.

(08:47):

Yeah, there’s always going to be an incentive, and even responsibility, almost, I would say, to capitalize on having a large user base. The fact that it is a foundational tool that enables really cool applications to work, and leverage the fact that there would be a high cost for those developers to rebuild things on a new platform. We see this all the time in big tech companies, for example, where companies like Facebook, YouTube, and a lot of these giants start up providing great services even at a loss to capture these really large user bases.

(09:19):

Once you have a large user base, it’s really easy to become profitable. Users are already invested in your platform. You can basically keep finding ways to profit from them as long as their switching costs basically outweigh the damage they’re doing to them. I think these ideas, these kind of patterns I was seeing in traditional business is what really initially attracted me to the idea of decentralized organizations, and how this model could be a lot better.

(09:48):

Just to quickly touch on another point, I think the other big piece for me here is that I think there’s a lot of potential around creating more equitable systems that function more like true meritocracies. I think why wouldn’t you want something like that? I don’t think web3 solves every problem in this aspect, but I do see it as a step forward from these traditional business models.

(10:10):

The very last thing I’ll say on this topic is that I’m also a really big believer in people being able to do what they love. I truly believe that you’ll be able to create more value if you’re able to follow your passions. I think web3 is also a clear enabler of that.

Nick (11:48):

Well, I love that answer, Ricky, and thank you for sharing that; a lot of philosophical points that I think listeners will understand and appreciate. I want to turn to you, Craig, now, because in contrast to a lot of what Ricky shared there, you worked for the big companies. You mentioned Google, you mentioned LinkedIn. You’ve got a PhD. How did you then find your way into web3 and crypto?

Craig Tutterow (12:20):

Yeah, it’s interesting. I would say yeah, I was kind of living in a parallel web2 timeline. My perspective might be kind of interesting for the OGs that have been around a while and saw the value from the beginning. I had been a crypto skeptic for a while. I had been following from the sidelines for a while. Yeah, I got into Bitcoin in 2013, and it was interesting conceptually, but then I didn’t see a broader use case for it. I kind of went away for two years and just focused on other things.

(12:56):

Then in 2015, there was an ICO for Auger, and Auger is a prediction market built on Ethereum. I was just interested in prediction markets because that was just an academic interest. That’s a good way to aggregate collective intelligence. If you’re buying insurance, it’s good to know what the likelihood of someone winning an election or something like that is. Those would always get shut down when they were on web2 rails. I thought, oh, it’d be interesting to have censorship resistant prediction market.

(13:29):

I didn’t really know much about Ethereum at the time. I thought it was a crowdfund sort of deal. Anyway, I just pitched in a little bit, just went away for about two years, didn’t follow anything. Then 2017, ICO boom happened. I came back and looked at the product, and tried to use it, and it was just completely unusable. It was just like a spinning wheel the whole time, would take minutes to load. I don’t know if they were using an archive notice on their database or what they were doing exactly, but it just did not work.

(14:01):

You could see all this hype, but on the ground, the products were not usable, and they were clearly not ready to scale to anything like what people were talking about at the time. Some skepticism set in there because I didn’t really see a path to scaling or mainstream usage. What I think the interesting thing is, that’s the parallel web2 timeline and crypto skeptic timeline, but people at the same time, the people at The Graph, the founders were looking at that exact same problem, and they were thinking about it in different way.

(14:36):

It wasn’t like, “Oh, this can’t scale as it is right now.” It’s like, “Okay, that’s a problem. We need to solve that before we scale.” Basically, they figured out a way to scale OLTP queries to be on par with what you would expect in a consumer web app by indexing blockchain data. That, to me, was pretty amazing. When I came back to it and made my first Uniswap transaction, the light bulb went off, and I started going down the rabbit hole. I figured, I have to figure out how is this stuff scaling now, whereas before, it was completely unusable?

(15:10):

That’s how I learned about The Graph. It’s kind of bridging the web3 architectures, and the web2 architectures, and making hybridization work there. That was very interesting. I was very impressed with the team, and yeah, just very excited to be able to work here and contribute to that, because what more could you want to do? How could you have a bigger impact on the world right now than working on this technology, and enabling what Ricky was talking about with respect to broader user ownership of platforms and technology and data?

Nick (15:41):

Yeah, I want to talk a lot more about The Graph and web3, and especially data and analytics. Before we do, it’d be helpful for guests to understand what it is you do at Edge & Node and what you’re working on. Let’s start with you, Craig.

Craig Tutterow (15:53):

Yeah, so far,, our work has mainly been setting up internal systems for monitoring and improving the network performance. We want to get the decentralized network, the quality of service up to really high standard availability and performance so that people can self-serve. We also want to monitor growth of developers, and make sure they’re having a good developer experience and so forth. Really coming up with these key KPIs, putting into place data pipelines to monitor those and dashboard those over time, and make those available within the core developer community.

(16:29):

Yeah, we haven’t even started to scratch the surface of what’s possible with web3 data science, and in terms of The Graph enabling the rest of the ecosystem. That’s something I really hope to get into more.

Ricky Eslapon (16:43):

The only thing I would say is kind of echoing what Craig said. We’re creating a lot of useful data-driven reports of sorts, which is something that I really enjoy doing and contributing to.

Nick (16:54):

The work you’re doing at The Graph’s obviously very important. It’s nice, because there’s a lot of charts, a lot of KPIs, a lot of data, and like you said Craig, there’s a lot more to come. It’s a very interesting time in the roles that you have. I want to go back to something you said earlier, Ricky, which was this, a lot of these philosophical reasons why you moved into crypto and into web3.

(17:15):

I’d be curious why your role at The Graph or why The Graph protocol itself addresses some of the concerns or interests that you had initially.

Ricky Eslapon (17:23):

The main thing I would say there is, again, as I mentioned, I’m creating a lot of these kind of data-driven tutorials that use live data. I was really fascinated by the idea of being able to create this permanent applications and permanent content. I think The Graph absolutely addresses these concerns. It’s huge, because I was actually, as I was building my project, I was running into a lot of issues around APIs that I was using, both becoming deprecated over time, as well as them having all sorts of rules and regulations around them.

(17:57):

Being able to use The Graph was actually a huge unlock for my project. I was actually starting to wonder, how am I going to be able to develop the projects that I have in mind? I was starting to think, wow, I kind of need something that works exactly like The Graph does today. I was honestly amazed that the team had thought of this as a problem three, four years ago at this point. That was amazing to me.

(18:20):

Yeah, I was very impressed by The Graph when I first came across it, and really enabled my projects to kind of go in the direction that I ultimately want them to go.

Craig Tutterow (18:28):

Yeah. I was interested in developer productivity. When I was looking at crypto and what’s happening here, the developer velocity is just, it’s insane, like the speed at which people are innovating and creating new things, and the extent to which things changed week to week. The other amazing thing is we were able to scale to this level with maybe one or 2,000 people building over a winter when everyone was kind of very skeptical that this would ever work. Now, we’re scaling to millions of users, and seeing this influx of developers and development activity.

(19:11):

The only way I can make sense of that is there is something to the technology to having a world-readable, world-writeable, a pinned-only database that everyone has access to with standards around open source code. Yeah, I feel like it’s a productivity multiplier in ways that we haven’t really seen in a long time. We’re also seeing a lot of experimentation around organization, how people are coordinating with each other in DAOs. One of the problems, macroeconomic problems has been the great stagnation.

(19:44):

Even though we’ve had all this innovation in terms of digital tooling, digital transformation, we haven’t seen that show up in the productivity statistics. One of my theories around that is that while the tools might be better, the way we organize ourselves is the exact same as it was 100, 150 years ago. There hasn’t been a lot of innovation in terms of corporate governance and organization. We’re starting to see that now, and that is very exciting. I feel like that will lead to some improvements in terms of productivity.

(20:15):

I think there is some sort of unlock that’s happening here, and The Graph is enabling that by making the data usable and queryable, because otherwise, it would be very difficult to access in an efficient way where you could use it in applications.

Ricky Eslapon (20:29):

Yeah. If I may just add, one thing that really also attracts me with The Graph is just the underlying fact that it uses GraphQL. To what Craig said, around all the great points that he just made, I think traditional APIs really only enable you to create data analysis and things around data that originate from how the API was designed.

(20:52):

You design an API with specific usage already in mind, versus GraphQL is much more powerful. It allows any user to use the infrastructure to create any kind of analysis and any kind of tooling or whatever around that endpoint, versus a traditional API, which, again, only allows you to do what the developer originally had in mind. I think that also really attracted me to The Graph.

Nick (21:16):

We’re talking a lot about revolutions that are happening right now. We’re talking about an evolution in the internet. We’re talking about a change in corporate governance and structure, modification of business models. When it comes to data and analytics, how would you explain that revolution that’s occurring, and how The Graph fits into all of that?

Craig Tutterow (21:37):

Yeah, the big difference in web3 is obviously that the data is public. Instead of having this kind of false choice between privacy and transparency, we can have both. By enabling pseudonymous transactions, you can have radical transparency in terms of the activity on the ledger, and have that be publicly visible to everyone. What that means is that the data isn’t locked into these silos. That creates a lot of really interesting opportunities and competition.

(22:08):

You don’t have as much lock in, because the data is available to any developers to piggyback on. Siloing data is the source of a lot of issues, if you think about in finance, being able to measure counterparty risk. That’s very difficult when data is siloed. If you think about moving from one social network to another, bootstrapping a social network, that’s very difficult when data is locked in and siloed. It gives users a lot more choice. It creates a lot more competition, and a more vibrant marketplace for software development, I think.

Ricky Eslapon (22:42):

Yeah, yeah, I totally agree with what Craig said. It really is about accessibility, because all of us in our daily lives, we use “data” to make decisions. Every decision we make, any action we take is always driven by data, even though it doesn’t come in the traditional sense that we’re talking about right now, pulling data out of an API. As people, we do formulate decisions around having information available to us.

(23:10):

I think making the data we’re talking about right now more accessible to the general public and anyone in a much more accessible fashion is going to just enable so many positive consequences from there.

Nick (23:23):

When we say data, I think what we’re talking about it here during this podcast, there’s probably two senses of the term. In one sense, it’s data related to The Graph, amount of users, amount of queries, and the types of reports you can run there. Then there’s this other sense of data, which is blockchain data, and smart contract data, and all the things happening on the chain.

(23:45):

How would you orient listeners towards the types of things you’re working on, in the two senses in which we’re talking about data?

Craig Tutterow (23:53):

Yeah, I don’t think we’ve even scratched the surface of what we’re going to be able to do with The Graph’s data. The Google analogy, looking from the outside, I was like, “Is that really accurate?” Just being an API, SaaS is like, for blockchain data, is big enough. That’s great. I wasn’t sure how much of a stretch that would be, but after learning more about the systems and being on the core development team for a little while, I think it’s definitely not a stretch at all.

(24:19):

It’s just we haven’t gotten to that use case yet. We’re trying to make that base use case of the API serving real time data. For applications, we’re trying to get that use case very solid first. In terms of the next few years, that is definitely a possibility where we see people doing more machine learning, data, mining, offline type data science analysis on blockchain data, using all the aggregated subgraphs. A subgraph is basically a lot of index blockchain data stored in a Postgres database.

(24:54):

Indexers combined have hundreds of terabytes of blockchain data. It’s going to be very useful in terms of coming up with new types of applications. If you think about a web3 social network, or a token recommendation engine, or a search engine, those are things that we don’t have any proof of concept for yet, but that data is all there in the Postgrads databases to do things like that. That is definitely something we want.

(25:28):

It’s on our long-term roadmap to basically enable people to be able to run these type of models, and to decentralize every part of the data science stack, so that The Graph is not only serving these online API use cases, but also what Databricks or Snowflake does for data warehousing in web2, having The Graph be able to enable those types of use cases as well.

Ricky Eslapon (25:52):

Yeah. Yeah, I completely agree with what Craig just said, and I would almost even go another step forward, and thinking about how incredible would all the things that Craig just said when you apply them to all public data? I think that could have a huge positive impact. It does make the most sense in the context of blockchain data to do something like this. Blockchain data has verifiable cryptographic proofs that essentially allows you to get to an objective truth about something.

(26:18):

Even going further than the three to five-year timeline, if you were able to do something like this on all public data, and do all the things that Craig just mentioned, that would have such a huge positive impact.

Nick (26:30):

It seems to me web3 is fundamentally data-based. It just seems like the whole foundation to use what you just said there, Ricky, is proofs, and data, and transactions. I’m wondering, with this ocean of data that seems to be created probably every second, how important a solution like The Graph is for web3 to really deliver on its promise in the world?

Craig Tutterow (26:56):

Yeah, it is critical, like I was saying with the Auger use case, all of the applications now are hybrid architecture, so a lot is still running partially on web2 rails. The reason for that is just because there’s a limited amount of block space. Not everything is going to happen on blockchains, but the most valuable transactions and interactions will. More and more will be happening on blockchains. The Graph is critical for all of those, and bridging web2 infrastructure and web3 infrastructure.

(27:33):

We have to enable some sort of hybridization in an intelligent way that preserves the benefits, the decentralization benefits of blockchains, but also enables their performance and all the things that have been built up to scale to a worldwide user base.

Ricky Eslapon (27:51):

Yeah, I completely agree. I know without something like The Graph, as Craig just mentioned, you do fall back into having to rely on web2 type tools. Unfortunately, it’s not very simple at all to actually extract the data out of the blockchains. You literally need to download the entire blockchain, and then parse out the data, and then figure out how to actually put it in a more usable format. You should also consider, how difficult would it be if you have to do this for every single blockchain?

(28:16):

We always talk about a multi-blockchain future. Having to do a different process for each blockchain and application you’re interacting with is really not ideal. Having something like The Graph which unifies all these things, and maintains those web3 ideals, like Craig just mentioned, is incredibly vital.

Nick (29:00):

Well, I really looked forward to the opportunity speaking with both of you, because I know you’re working on data at The Graph, and you get to see and observe things that a lot of us don’t get to. I’d just be curious to open it up a little bit and get a sense of some of the things you’ve seen or observed that have surprised you or interested you.

Ricky Eslapon (29:21):

Yeah, without getting into anything super specific here, I do think that it actually has been surprising to me, just the speed and scale at which The Graph as a protocol is actually ramped up to something functional. I think there’s a difference between the theory of something working the way you would expect it to, and then something actually working that way.

(29:42):

I think in theory, it’s amazing to think about how The Graph could actually be more reliable than traditional databases, and faster, and all these kind of amazing attributes around it, and actually be able to see those kind of improve over time and already getting to very, very good levels of quality of service for the users has been really incredible to see.

Craig Tutterow (30:04):

Yeah, it is amazing to see the amount of throughput and the amount of data that’s being generated, and also the quality of data just coming in. In terms of queries, The Graph has very unique position in that it’s a gateway to a lot of different web3 sectors, so whether it’s gaming, Metaverse, Defi, DAOs, they all have subgraphs. Yeah, the scale has been amazing. Yeah, we want to open source more of that data and get it out there, so that people in the community can use it to power their decisions.

(30:38):

Whether it’s quality of service on specific subgraphs, so people are getting real time updates on that subgraph, developers are able to see that real time, or Indexers knowing how the Indexer selection algorithm works, and where they’re ranking within that. Those are definitely things we want to be able to push out insights to network participants in real time. Then also stuff like business intelligence, what percent of web3 users are mobile in different sectors, or what’s the Geosplit?

(31:09):

Ricky has mentioned doing some deep dives on Geosplit across chains and things like that. This sort of usage data I think is going to be really interesting, and that’s also something we’re looking to open up and work with the community on publishing some of those insights. Yeah, those are the things that have struck me.

Nick (31:29):

Can you give us a specific example, Craig, of an insight or something you’re working on to that point?

Craig Tutterow (31:35):

Yeah, so I can give one. Since we’ve been focused internally, I can give an example on the decentralized network, some of the deep dives we’ve been doing. I think most people in DevOps and most site reliability engineers would understand intuitively, that outages or failed queries are a bursty thing. You can have an outage, usually, things will be performing great, four nines, et cetera, and then there will be an outage, and it’ll dip quite a bit, but very briefly. Then you recover from the outage and so forth.

(32:08):

Performance can be very bursty. What we noticed in the Indexer selection algorithm is that it was routing queries to the Indexers that were performing the best, of course, but sometimes, those high performing Indexers would get overloaded or have an outage, and the algorithm wasn’t responding quickly enough to those outages. It was like their score was decaying kind of slowly.

(32:31):

We found that if it responded more quickly to an outage, and treated those qualitatively different than something where their performance was just decaying linearly, that we could improve quality service on the network quite a bit. We could reduce the number of errors like 30, 40%. That’s one example of a deep dive that we’ve done where we identified something that could improve network performance, and we’re doing more of those.

Nick (32:56):

Ricky, when people get involved with The Graph, they’ve got a lot of things to learn. They have to reorient themselves to what an Indexer is, how token economics work, and a whole host of other things about the web3 and crypto space in general. When it comes to data, how do people need to approach or reorient themselves to that in a web3 type environment?

Ricky Eslapon (33:18):

I actually think that’s one of the beauties of The Graph is yes, there is a lot to learn in terms of the roles, but I would also say that if you really dive deep into it, it’s really necessary. There’s no kind of added tokenomics just to have tokenomics. I think everything that The Graph has done from the roles perspective is really necessary. Once you familiarize yourself with those aspects, I think as a user who’s interfacing with this data, there’s really no difference in The GraphQL endpoint that you’re using, whether it’s a traditional GraphQL endpoint or one that comes from The Graph.

(33:51):

I think the team’s done an excellent job in that sense. I’ve been really happy with that, in that transition of just a regular user, to coming into being more deeply involved with The Graph. I think those experiences are actually quite similar. It’s just the nature of the data is a bit different, obviously, but as a user experience, I would say it’s actually quite similar, which I think is definitely the approach to go with.

Nick (34:13):

Craig, given what Ricky just explained, I’d be curious, given your background at LinkedIn and at Google, obviously, two organizations with a lot of data, and I would assume a lot of viewpoints about the future of data, what is it that maybe they need to better appreciate, or understand? Maybe speaking to yourself seven years ago, what you wish you would’ve better understood about blockchain and the future of data?

Craig Tutterow (34:38):

Yeah, my recommendation would be to not think about blockchain or web3 is an all or nothing thing, where either everything is going to be on the blockchain and run on blockchains, versus we have to have entirely centralized siloed databases within these large monopolies. It is possible to have an intelligently constructed hybrid architecture, where some things are stored on the blockchain and other things are happening more in derived databases like what we do in The Graph.

(35:11):

I think that people in web2 would probably be skeptical that web3 or blockchain systems can scale to accommodate the type of user base that they work with, which is billions of people. I think that’s a valid concern, but that has been massively de-risked, I think, partially due to The Graph. Yeah, so I think there’s a very optimistic story to tell in terms of scaling with some sort of hybrid architecture.

(35:36):

I do think many people are very sympathetic to the ideas of web3. I don’t think most people, most software engineers in web2 got into it really wanting to monopolize data collection or anything like that. I think people are very sympathetic to the ideals, and you see that with a lot of founders pivoting their companies to more of a web3 ethos. They’re questioning how things have played out in web2, and thinking about how we can fix some of these things. I think there is a lot of sympathy there within web2.

(36:12):

Yeah, it’s just a matter of thinking about, how do you solve the problems? When you see problems, don’t see it as a fixed thing that can’t be resolved. This field is moving very quickly, and there’s a lot of optimists. It’s just not a zero-sum kind of mentality, where things are built, and you’re kind of fighting over a fixed pie. Things are wide open, and you have the opportunity to come in, and contribute, and really shape the way things are being built, and make big contributions.

(36:43):

Yeah, I would just encourage people to look more into things and figure out how they can contribute.

Ricky Eslapon (36:47):

Yeah, I totally agree with all the points that Craig just made, and I agree. It’s about the sustainability of your company. This stuff is coming, and businesses, they just have to adapt all the time. You’re probably going to be out competed if you don’t adapt in some way. I think one mistake that I see, my old colleagues reaching out to me asking me about blockchain and how they can integrate it within their own companies, they see it as like, “Oh, this is just like AI. We’re just going to plug it and play it into our current systems, and we’re just going to make more money,” type of thing.

(37:19):

That’s not the reality of the blockchain. You do need to fully understand what it does, why it does it, and transition to that model a little bit more. Yeah, I totally agree with the points Craig made there.

Stefan (37:40):

Hey, this is Stefan from The Graph Academy, announcing The Graph Academy Grants Program, a program that is intended to foster the growth of our community-driven platform, The Graph Academy. We are offering grants to contributors to incentivize the creation of high quality educational material. If you want to apply, simply visit TheGraph.academy/grants. That is TheGraph.academy/grants. You can also find more information in the show notes.

Nick (38:11):

Well, this idea that web3 isn’t an all or nothing, or either or type of scenario, meaning that blockchain and web3 can exist, while some features or organizations in web2 exist, has come up multiple times on the podcast. How would you help listeners better understand this type of dynamic, where you have a web3 and a web2 coexisting together? For me, still learning these things, it doesn’t seem like it’s possible.

Ricky Eslapon (38:53):

Well, one thing I would say is, to me, it really is about progressive decentralization. If you do decentralize something too soon and give all that power away to this automated web entity, there’s a lot of risk that is associated with that, especially as we’re still trying to do a lot of work around things like decentralized identities. There’s a lot more ways in which you can attack these types of systems.

(39:15):

I think for the foreseeable future, there’s definitely going to be for sure some kind of coexisting between the two for that reason. The long-term, this is just kind of my personal opinion, but I would expect web3 to just be a more efficient, better way of a lot of current systems. When people kind of act like, “Why do we have to blockchain this and blockchain that,” there’s very little that I could point to that wouldn’t benefit in some way from moving more in that direction.

Craig Tutterow (39:44):

Yeah. Some things like CloudFlare are just not going to go away in SQL databases. To scale to billions of people, you’re going to need those things. I guess what we’re all calling web2, there’s the business models, and then there’s the technologies. The technologies underneath, those are going to be needed, no question. They can be implemented in a more decentralized way, which is what The Graph is doing.

(40:11):

We’re using Postgres databases, a lot of open source software, and these systems that can scale under the hood, but we’re enabling community ownership and participation by Indexers who are running the infrastructure, and getting the Indexer and query rewards, and Curators who also participate, and Delegators who also participate. The underlying technologies that have been established to scale web2 systems, they’re still going to be around for sure.

(40:40):

In that sense, probably the cloud provider’s still going to be around, but maybe the business models will get disrupted. I just think there are massive incentives and a very strong value proposition for as a user to be able to share in the upside of the platforms that you use. Like Ricky was saying earlier, instead of the benefits just going to a small group of shareholders, users, they have some ownership, and some say in the protocol development or platform product development.

(41:12):

That’s just, I think, a very powerful value proposition right now.

Ricky Eslapon (41:16):

Yeah, and one more thing I would also say is I feel like people have this notion that web3 and web2 are so fundamentally different from a person working within a company, but in reality, I think it’s just about the structure of the company. As an analyst, whatever your job is, your experience is going to be similar when you’re interacting with a web3 protocol. Right now, we’re all talking to each other, we’re all in traditional roles, even though we’re working for a protocol.

(41:45):

It’s not like there’s everything kind of functions as an open source DAO for every little decision. You kind of delegate things to individuals in similar ways that you would in web2. You just do it kind of aligning the incentives differently.

Nick (42:01):

Well, I appreciate both your ideas on this interesting dichotomy between web2 and web3. If we think about the web3 stack, about the things that must be in place for web3 to fully be actualized, what, Ricky, would be part of that web3 stack to have at least something, or a framework, or rails upon which web3 to fully be adopted?

Ricky Eslapon (42:24):

One thing that I would say that comes to mind is something like GitHub, that is more kind of open source and decentralized. We do have Radical, but we always hear about a lot of problems with Radical. That’s one thing that comes to mind. GitHub is obviously where everyone collaborates today, and it was bought out by Microsoft. That’s one piece.

(42:44):

One other thing that really kind of jumps out to me is more related to the work I do around the Metaverse, and this idea I have of creating a public open source university in the Metaverse. I think a lot of the hardware side really needs to catch up. I think we are kind at risk of bigger companies like Facebook, Apple really being able to get ahead and dominating things on the hardware side, which is really difficult for us to solve.

(43:13):

In terms of software and protocols, I think something for code management like Radical probably is a big one. I’m sure Craig has some other thoughts in this area.

Craig Tutterow (43:24):

Yeah, I think we’re talking about something that’s going to have to happen on a decade or multi-decade timescale. Yeah. If you look at most of the coordination that’s happening, it’s happening on Discord, Twitter, LikeSo Chat, things like that, we would need decentralized applications for that. I do think that’s how things will trend, but I do think it’s going to be on a longer time horizon probably than people are thinking right now.

Nick (43:53):

Craig, I want to come back to something we alluded to a little bit earlier, but now in the context of this non-binary situation between web2 having to lose for web3 to be fully adopted, that these things can coexist in some ways; and addition to that, this timeline, that it’s going to take more time, more build for web3 to be fully evolved. In that context, what is then your long-term vision for The Graph?

Craig Tutterow (44:20):

Yeah, I think long-term, I would like to see The Graph decentralize every part of the data science stack, and empower anyone in the world to use blockchain data to develop applications. Right now, we’re kind of focused on the OLTP use case with live transaction data streams, but in the near term, I think we’ll have some more support for OAP type queries, like aggregations.

(44:46):

I also want us to be able to support that decentralized Databricks or Snowflake use case, where people are able to query on really large data sets and do joins across multiple subgraphs, like terabytes of historical data, and fit models on user behavior to better understand how people are using things, and then also produce recommendations and enable web3, social media, or search engine use cases, where people can bring their own model to different environments.

(45:23):

Some people might want a model, like you hear a lot of controversy about social media amplifying controversial takes, and creating more antagonism in society, more conflict by amplifying those voices, in order to produce more engagement, which is beneficial to their business models. Maybe some people, that’s not the type of content they want to see. People should be able to choose, A, what type of data the model’s trained on, so can you use my data or not? B, what type of content do I want to see, or what type of app environment do I want to be in?

(46:03):

Maybe there’s a super civil social network, where content that is trying to bridge across political divides, that’s the type of content that gets amplified. The discussions on that platform are going to be very different from the discussions somewhere else, where the most salacious stuff is getting amplified. Giving users the ability to opt in to what type of data is being used to train the model, and also giving people more control. I think those are the types of use cases I would like to see The Graph empower in the future.

Nick (46:38):

How about you, Ricky? What’s your long-term vision for The Graph?

Ricky Eslapon (46:41):

Yeah, I think Craig actually did a really great job summarizing that previously and just right now. In fact, the example I was going to bring up is exactly the one he just brought up of the bring your own algorithm to social media or whatever website you’re interacting with. That’s actually the exact same example I had in mind.

(46:58):

I think he’s already done a great job covering that. I think the other one that we also kind of touched upon earlier is the public data use case, and how we can expand this framework to all data under these same rules.

Nick (47:12):

Craig and Ricky, thank you so much for your time. It’s been very interesting to hear more about your roles in data and web3 from your perspective. If listeners want to learn more about some of the work you’re doing, what are some resources or some things that they can have access to to better educate themselves?

Craig Tutterow (47:29):

Yeah. Our focus this year has been on powering internal decision making. We’ve been setting up internal systems for monitoring, and metrics, and dashboarding, and so on, to get the decentralized network quality service really strong and ready for self-service onboarding. Our focus in the new year is going to be on engaging the community and open sourcing more of our data to power external decision making, not only of participants in The Graph ecosystem, but also web3 in general.

(47:56):

In setting up our internal systems, we’ve basically been dogfooding state-of-the-art tools, and learning a lot about the trade-offs with an eye towards designing a full data science stack for processing on off-chain data. Our data engineer, Aaron, has done an amazing job setting up our data warehouse environment and establishing the pipelines that power our analyses.

(48:17):

He’s going to have a blog post coming out soon to discuss the streaming architecture he’s set up for real time data processing. That’s going to power a lot of the dashboards on The Graph Explorer and some other run-in contexts.

Ricky Eslapon (48:33):

Yeah, definitely more to come on that front. I would definitely point people, again, to the blog posts. I think there’s been some really great posts, so the ones by Brandon Ramirez, around protocol economics, and the deeper dives that have been taken on the blog posts, I think those are really great resources for users. I also think community members, like Graphtronauts, have done a really great job. Again, he was also on the podcast, so recommend users to check that out for sure.

(48:59):

Then kind of on a more personal note, I’ll also be working on this concept of an open source, decentralized university and the Metaverse, which will be all powered by The Graph, which will take deeper dives into real problems, real topics, and hopefully be, in the longer term, be a great resource.