Trent McConaghy, is not only one of the early blockchain entrepreneurs founding several startups in the space, most recently Ocean Protocol, a decentralised data marketplace that enables a New Data Economy but an AI founder who enjoyed a big exit. He explains how Ocean can break down entrenched data silos across industries and allows for the gains of Big Data and AI to be more evenly distributed across society.
As the founder of the global token engineering community we talk about why the principles of ‘engineering’ are critical to building safe infrastructure for a new multi billion dollar digital economy centered on Web 3.
- The challenges of the data economy that exists today
- What a New Data Economy looks like – both today and in the future
- Why we need data in DeFi
- The rise of decentralisation – and what is going to be needed for crypto has to be big enough and mature enough to handle it.
- Thinking about token engineering are the combination of theory, practice, tools and responsibility for designing, deploying and maintaining
Jamie: Welcome to the Founders of Web 3 series by Outlier Ventures and me your host Jamie Burke. Together we’re going to meet the entrepreneurs the backers and the leading policymakers that are shaping Web 3. Together we’re going to try to define what is Web 3, explore its nuances and understand the mission and purpose that drive its founders. If you enjoy what you hear, please do subscribe, rate and share your feedback to help us reach as many people as possible with the important mission that is Web 3.
Welcome Trent. Good to have you on the podcast. You’re one of the first people that we´ll be interviewing I guess in a way you’re my comfort blanket like my go to interviewee. I think we did a podcast at some point in the past and we s o I had you on that. And that’s for a number of reasons. Firstly, because we know each other quite well, we’ve collaborated in different ways. I think we share some common ideas and themes, and I just really like talking to you. So any opportunity to do that I kind of jump on. So welcome to the to the show, I guess we’re going to call it.
Trent: Sure. Thank you. It´s a pleasure to be here.
Jamie: So as I explained to you before you came on, what I want to try to do with this podcast and format is to cut out all the fluffy stuff that you would normally do the intro to a podcast, you know, you’re fairly well known in the industry now. And you’ve also been on a lot of podcasts, you’ve spoken a lot of events. So I’m sure people are familiar with you and your background. And so what I really want to try to do is to explore some of the themes that define you the meta themes that define you and Ocean and your role within that. And then also explore some of your fresher thinking that might be a bit unformed and actually bounce some of the thinking that I’m having, we’re having an Outlier. So, but it would be good to kind of get just maybe 30 second intro for those of you that that won’t know you.
Trent: Sure. So I guess the quick 30 second intro is that I’ve, you know, was raised in rural Canada, have done two start-ups in the past, both in a AI around designing computer chips. And I guess for since 2013, I’ve been working very intensively on blockchain and more recently with blockchain and AI, most notably Ocean Protocol. So I guess that’s the quick summary.
Jamie: Yeah, that’s exactly what we needed. And I think you know, one of the really interesting things about you and the time that I’ve known you, as we’ve been kind of defining our thesis and understanding the landscape is, that you are one of the first people from AI that was crossing over into the blockchain space. And clearly that’s gone on to inform Ocean now. But it would be good to understand from you why AI and blockchain or DLT for you are perfect bedfellows, you know, what, what was it that helped you make that connection? And why do you think both of those things in combination are important?
Trent: Right. So there’s probably two levels, the broader level on the soil level, the broader level is that AI is one of these very few technologies in the world. That has is a general purpose technology that can have potentially big impact. And it’s not really fully exploited yet. So it’s a really great lever for the world to help change the world for the better; blockchain is also one of those. And you know, there’s a few others things like CRISPR and all this but blockchain AI are much more it related, and that’s really what I know best. So that’s probably why I have gravitated towards both very much. So that’s kind of the broader. If you drill down a little bit more specifically, with how they intersect, there’s quite a few different intersections. I first wrote about this several years ago now, but the main intersection is that modern AI in the form of deep learning, you know, these big, big neural networks, modern AI loves data. And so, if you’re trying to build an AI application that is accurate enough to ship and make money on, you know, you kind of an error rate of, say, 30%, you know, it’s just going to be infeasible, you need, you know, 10% or 1%, or 0.1%, or whatever it is for your application. And the traditional way to do that was to spend, you know, two years five years doing a PhD to improve an algorithm for this. But since the mid 2000s, more and more and more AI researchers have realised, what you can do is simply add more data. So instead of being a computer science mystery, it’s more of an engineering problem. So by just adding more data, you can improve the accuracy, you know, from an error rate from say, 30% to 10, down to say, 0.1% or whatever. Just to simplify Hearing cask. And so then the question is, where do you get the data from and so on. And it turns out that a lot of the world’s most valuable data, the world’s most useful data is private, because the people holding it want to keep it private, for privacy reasons, or for reasons of control. So that’s, you know, the intersection is that blockchain can really help to unlock data while maintaining privacy.
Jamie: Right. And, and so I saw your recent presentation in The Hague, which kind of did a really good job of kind of qualifying and quantifying the impact of data on machine learning models, but it isn’t a given right. There is a, it is a contentious point, I have heard some people say that, you know, just increasing the volume of data doesn’t necessarily mean to better outcomes, or it does but only in certain fields.
Trent: So it’s well though there can be dimensional returns but usually the people that are saying that have have a hidden agenda, like for example, something that say AI researchers, they want to promote their work on symbolic reasoning. And then work is really good. They’re really high quality researchers. But that’s not to discount the huge impact that simply adding more data is.
Jamie: And it’s one of the arguments I’ve heard is that, you know, more data works if it’s in a very narrow form of AI. So I think the use case I heard was around machine vision and in the context of autonomous vehicles, but it’s less relevant. If you’re kind of trying to move towards something that’s a bit more generalised. Is that a fair summary? Or do you can kind of contest them.
Trent: I think overall, AI is basically a group of different fields, right? You’ve got fields like neural networks, evolutionary computation, symbolic reasoning, all of these and, and then they’re solving problems that have often traditionally been computer science mysteries that can’t really be viewed in a very, you know, pure symbolic sense or pure numerical sense. And, you know, these AI tools, solve them in novel ways that are, you know, less direct, if you will, sort of more soft computing type ideas. And so within those, what’s really taken off in what’s sort of part of the Zeitgeist of modern AI is neural networks that are really, really big, right, deep learning all that. And they’re solving just very specific problems in regression and classification. And those are, the applications for those are large. Initially, it was a lot of computer vision applications. And more recently, we’re also seeing a lot of natural language processing applications. So that’s, you know, when we hear about start-ups and the AI Zeitgeist, more than 95% of the applications are applying deep learning to typically on one or both of these applications. However, that is just one small slice of the overall field of AI compared to all the other possible, you know, AI tools and techniques that are out there. And a lot of these other tools and techniques don’t need to data as much, right? If you’re not doing regression or classification, which is really about building models, where a lot of data can be valuable, then it’s less useful. You know, for example, what I did in the world of circuits, we were building models, but we only had access to circuit simulators. And we were very limited in the budget, that we could have to get data from the circuit simulators. So we had to make do with, with the data that we had. And so that wasn’t, you know, the start-up wasn’t more, couldn’t be more successful by simply adding more simulations, because that was, you know, a very budgeted. So we did have to find ways to, you know, make better use of the data. And there’s other AI technologies out there that are more about just simply learning from the environment. Think of the world as one big infinite data set. So how do you go around and explore that world, whether it be the actual physical world, or some virtual world? So overall, it does depend, you know, a lot of AI techniques. Don’t think about data, they don’t need data at all. But the AI that’s in the Zeitgeist, that’s really you know, having these apps a lot of low hanging fruit are mostly data intensive.
Jamie: Right and you know, I think, increasingly an Outlier, we refer to what we invest in and the way that we look at Web 3. And we’ll get to that a little bit later as a new data economy, enabling a new data economy enabling for the commodification of data. And a lot of that thinking has been informed by you know, your work at Ocean. And I know that if you look at how you’ve constructed the economy that will form around the Ocean Protocol, yet, an increase in the volume of data kind of unlocking that long term effect is just one part of it, right? You also have availability, timeliness, this idea of quality that you can curate, and maybe you could talk to some of those points as well.
Trent: Sure. So overall, maybe starting with the macro lens really more deeply. We, there’s actually already a massive data economy out there, but it’s shadow data economy, so there are, you know, Bloomberg, for example, he became a billionaire by selling these Bloomberg terminals to traders in Wall Street going back to the 80s. Right, you know, got famous enough to go into politics and even run for president right? and and you know, and buying his way for the ads. So that’s one example. And so Wall Street actually has been using data, bind data and using it for decades, right? So that’s a vertical. And other vertical that uses data like crazy is social media. So Facebook is the most famous example there, and they use it, you know, they get people to enter data to connect to their friends and stuff, but then they use it in order to sell ads. And so you can generalise social media into sort of the ads industry, where it’s really Google and Facebook is the leaders there. And Google and Facebook themselves, they’re using data, they’re using data that they gather, but they also buy data from a lot of other agencies. You know, last time I checked, Facebook was buying from more than 150 agencies just in order to sell ads better and you can go vertical direction. To vertical, and you’ll see these sort of mini economies. And they’re not small, like each one is big on its own, but they’re fragmented. So you can go, you can look at healthcare, you can look at logistics and transport, etc, etc, etc. And if you sum it all up, it’s already massive actually. And it’s already, according to the World Bank, several percent of the GDP of the world, I forget the exact numbers, but a few percent anyway, and it’s rising all the time. And so the data economy exists, but it’s hidden, we don’t see what’s going on. It’s sort of sold quietly by brokers, and consumed quietly by people who don’t necessarily want to know that it’s often your data being bought and sold, such as in the case of Facebook. So that’s why we think of it as a shadow data economy. There’s a lot of money flowing around, but it’s not serving the best interests, especially of the public and consumers. And, you know, just like the idea of Bitcoin and the broader blockchain industry initially and continuing has been about going from a shadow banking economy or shadow money economy to an open permissionless money economy. That’s what our aim with ocean is, is to go from a shadow data economy to an open permissionless data economy, one where people work that does reconcile privacy, you know, where people are Shades of Grey, a controller privacy. So that’s kind of the macro level. Maybe I’ll stop to see if you have questions before I drill into sort of like, if you want I can go into the more detailed interactions among buyers and sellers and stuff.
Jamie: Well, I think what would be good is to kind of make the connectors to why this is now possible as a consequence of DLT and blockchain.
Trent: Yeah. So yeah, happy to. So overall, you people have been buying and selling data with data marketplaces in the past, but not in a big way. Usually, it’s a data is gathered kind of quietly and sold directly by the people doing the gathering to the consumers. So there’s not a lot of sort of more open ish data marketplaces. And they’re, like mentioned before, the two biggest reasons Privacy and control where the people doing the selling, don’t really want it to expose, you know, who they’re selling to and what they’re selling in any big way. So they don’t really want to put in the data marketplace. And in centralised data marketplace poses a lot of issues. One of them is custody, right. So just like, you know, a centralised token exchange, we have decentralised exchanges, the centralised token exchanges have got hacked time and time again, starting with Mt. Gox. And continuing to this day, even by Nance and big hacks, right. And so if the, whereas decentralised exchanges, one of the big things, they don’t control the keys of people’s tokens. So their attack surface is therefore much, much smaller. So that’s a big thing. The other thing is, and that’s what that’s one of the things that blockchain helps to lock in is basically, you know, addressing this custody thing, because if you have a decentralised data marketplace, then that middleman that’s connecting buyers and sellers isn’t holding liability, they are holding keys for the data and so on. They’re just basically connecting the buyers with the sellers and then the buyer and seller interacting directly with each other.
Jamie: So maybe to kind of just put pause on that, that the fact that we built all this infrastructure to take custody of digital assets is now all kind of transferable into when that asset can become your data. So I know this concept of bring your own data, but primarily what we’re talking about is using this public key infrastructure and ledger technology in order to take custody of and exchange this, this new form of digital asset.
Trent: Yeah, exactly. So you know, we think of each data set as its own asset and then also the typically that’s held by a publisher or a data publisher. Sometimes they’re called Data Brokers. And they can, you know, sell those data sets, as many they what they’re selling is access. So, from an IP perspective, they have copyright and then they’re selling licences. To others in order to licence to use that, that IP, in whatever way, right, you know, maybe for consumption in order to train AI models, maybe other reasons. So that’s kind of, roughly speaking, a way of thinking about it. There’s these, you know, data publishers, there’s data buyers, the middle men in between are the data marketplaces, right? And then there can be referrers, and all that too, and many other actors, but that’s kind of the core. But you know, going back then to sort of the USP of how, what blockchain brings to the table. So I’ve talked about custody, and that can really improve the data marketplace side, I hinted at the controlling privacy, but I can be more specific here. So what you can do is you can sell access to a data set, but you’re not actually selling, you know, copies the data itself. If you don’t want instead, you can make it where you’re selling such that the buyer can bring his AI training algorithms, right to the data itself. So, and by doing that, there’s human eyes never ever look at the data set, it’s for AI eyes only. And this actually resolved the issues of privacy and control. Because the person who’s owning the data, who’s published it initially all that they’ve actually never, you know, it’s never downloadable per se. Instead, this AI model is being calculated right next to the data itself. So the data never leaves the premises, the AI model is calculated, and then it’s used in order to, you know, for whatever value the data buyer has for the AI model. So that’s really, you know, really solve this problem, because of sort of privacy control right? Before it’s like, okay, you can either not make the data available, and then no one gets the benefits of more accurate models. But also you don’t have concerns about privacy control, or you can make it available and then have those privacy control concerns. But by doing this by bringing compute to the data, then you can have your cake and eat it too. if you will. You get more data for better models, as well as address the privacy and control issues and you know, this is with simply like compute containers living, living next to the data itself. There’s other ways to approach privacy as well that are, you know, coming down the pipe, they’re not quite as mature yet as just good old docker containers, things like secure enclaves, homomorphic encryption, all that. And we’re hopeful about, you know, those technologies as they mature going forward. And, you know, we will be helping to incorporate those sorts of technologies into Ocean with, you know, other companies as they develop them. But our first focus is the computer data.
Jamie: Right. And so, most people, I’d imagine is that kind of remotely tracking what’s going on in the web will be aware, Tim Berners-Lee has been recently rallying against, I guess, all of the ills that you’ve been speaking of, in the context of data economy and privacy and data custody, and obviously, he’s taking a very specific approach with pods. We actually did a review on, Amazon’s CTO and CEO, one of our technical analysts did a comparison of looking at sovereign, for example, in an SSI context, and Tim Berners Lee is solid, but I don’t know, do you think pods and what they’re trying to achieve that is complementary, or it is it is an entirely different approach to what you hope to come about from when users and that could obviously be a corporation or an individual begin to take custody of their own data.
Trent: So I think it’s mostly complimentary. And, you know, obviously, I greatly respect the work of Tim Berners-Lee, going back, you know, for decades and stuff, of course, and so it’s mostly complimentary. I think what you know, Ocean is is focusing on is a different aspect, so that, you know, where there are opinions on how to do self sovereign identity and have personal data storage, you know, it can be a sovereign type approach, it can be a solid slash pods approach. And all of those can plug into Oceans, Ocean is not trying to be too opinionated about these other, you know, complimentary problems to solve. Right. And so this is why Ocean is focusing on really the the data marketplaces for price data as well as for the Data Commons. And then the exchange of value around that. Right. And so kind of related, right, I’ve talked about custody, which is really, you can think of custody as, think of it like a crypto wallet where you, but instead of having just custody of crypto tokens, traditional ones, you have custody of datasets. I’ve talked about the, cracking this problem of privacy and control by bringing compute to the data. There’s a third one that you kind of hinted at. And it’s worth mentioning here because it also relates to your question with Tim, and that is data tokens. So it’s something that we’re working on quite aggressively now. Where, up till now, we’ve been leveraging Web 3 infrastructure in terms of private keys and the wallet, you know, managing private keys. But we said, hey, what if at given sort of lightweight token, a data token is simply an access control to be able to download a dataset, or to be able to get access to streaming data over time, or to be able to bring compute to a data set, any one of those three things, right static data, dynamic data, and compute for private data, basically. So imagine where I’m a publisher, I’ve got, you know, my data that I’ve created, maybe some location data. And then I mint a data token for call it location data token. And like basically deploying an aetherium contract. And then from that, I might mint 20 tokens from that, 20 type tokens. And then with those 20 tokens, I can be holding those in my crypto wallet, even meta mask or something, but then I’ll also I can be sending them to data marketplaces for those days. And marketplaces to be buying and selling. And then others can be bought, can be buying from there and exchanging and trading and so on. So then you’ve basically got, you know, any degree of fungibility you want, I can make 20 tokens, I can make 2 million if I want. And then people can be buying and selling this. So it’s removing friction for buying and selling access to data sets in any one of these three forms static dynamic or the private type data. And by doing that, you know, going back to your question with Tim and stuff, and solid, it’s they, people who have their the personal data can choose to make aspects of that personal data available for consuming by others. And you only need three forms, and they can do it by minting these data tokens. And so it’s a pretty exciting concept. And, you know, we’re in the process of building and rolling it out as part of the overall Ocean Protocol offer.
Jamie: So do you think you know, obviously, the way that we pay for the web at the moment I say we as a general user, we implicitly pay for the web and the majority of services with our data. Now, do you think that will continue? Is that the kind of, do you think that model will continue, except it’ll just become more explicit, and therefore, you’ll have more as the data custodian, you will have more control over the terms of, of that transaction and presumably revoke ability.
Trent: So two parts, I do see that as time goes on, people will be able to have much better, more fine grained control over their personal data, probably it will probably be look a lot like, you know, if you open your iPhone to the settings, you can see, you can swipe, what sort of permissions you give each app. And so you can think of it like an extension to that. That’s one way. Another way to think about it is you know, you’ve got your crypto wallets on your phone and in that inside that crypto wallet right now, ether and Bitcoin, but in the future, you might have access to you know, 1000 different datasets and Some of those are yours and you can send those others and so on. So that can happen. But of course, it has to be a very, very easy user experience for consumers, and it has to be worth their while in the first place. So I, we see that, in general, there’s a lot of directions this can go, we want to make it really easy for people to experiment with different ways to sort of capture value in this Web 3 data economy. So the stuff around sovereign personal data might be one way sort of consumer level, right? And, and it also might be around, you know, small and medium sized enterprises. For example, we’re working with X ray for building a data marketplace around logistics, you know, trucks and so on thousands of trucks. And or it can be, you know, at the level of large enterprise to and we work with large enterprises, or for their large data that they’re just mostly having sitting there gathering dust and even exposing them to liability in case of hacks or governments, cities, etc. So we don’t know which of these will pop first. So we’re making sure that what we’re building is general enough to use sort of across the board and an easy to use enough. And we’re not there fully yet, but we’re getting better and better and easier to use all the time. And then you know, where we see flames help to pour gasoline in those flames if you will to help catalyse things. And you know, but it doesn’t lock some exciting new possibilities, things like, you know, the radical exchange, folks, Gladwell, and all those folks there. They talk a lot about data trust data unions, and this points to the sort of personal data sovereignty as well. Well, with with ocean that can be a lot easier to implement, because you can have a Dow that’s implementing a data trust or to cooperate in a union. And then those Dows are simply holding Ocean generated data tokens. And so it just looks likes 20 token basically, and then they can be exchanged for value and this could be around say, you know, you could have 100,000 people buying and selling, you know, aggregating location data and then having people in the Dow working for the Dow selling their location on their behalf.
Jamie: Yeah, I mean, there’s two interesting parts of that. On the one hand, there’s the Dow element. And I know, you know, both of us have been very interested in this space of whether it’s a commons Co-Op, or a union where you kind of have collective or shared ownership of assets, including data. But one of the interesting things I just kind of want to push on a little bit. So we were talking earlier about the potential for this world where you would have a wallet, you would have custody of 10’s, hundreds, thousands of data tokens, obviously, if we look at parallels with the world as it is today, just in terms of holding crypto assets as they are the majority of people have deferred to a third party custodian in a very centralised way. How do you envisage the world of custodian’s agents in this context because you know having been engaged with a sovereign project for a very long time? You know that thinking about agent based systems and how custody can be delivered in a manageable way where the end product is going to be grandma proof as they say.
Trent: I our perspective, the notion is that there’s going to be shades of grey where from one level where the individuals or individual companies etc they are you know, managing taking all the effort to do the full control themselves and there should be really great UX around that or on the other end, you could have some centralised party playing a full custody role, the way that some people use Coinbase or by Nance and there can be Shades of Grey in between an important shade of grey in between is delegated custody. Okay. So, so just like, you know, an example is liquid democracy where you have delegated custody of voting incense. If you find people that you know, have spent way more time than you to, to research particular topics and then you delegate your votes. To them around these topics, right? delegated proof of stake for L zero mining as well, you know, I’ve got, well, in general, you know, the things like cosmos and so on, a lot of people, they don’t have atoms, but they don’t want to be running a node in the cosmos network. So they are just delegating to the existing nodes. And of course, there’s only a limit on the number of nodes that actually earn fees Anyway, there. So that’s, you know, an example of delegation. But the key thing with liquid democracy or delegated proof of stake is that at anytime, you can revoke your permission, your delegation, and you can control it yourself or reassign it to someone else. And this is the critical part, because it gives you optionality. And it means that the people that you’ve delegated to, they have to stay on their toes, and they have to continually provide you good service. Otherwise, you know, you walk away with your data tokens with your dollars basically.
Jamie: So that doesn’t it, so in that world, there’s a spectrum. It doesn’t preclude a Google or a Facebook adopting this new paradigm. But what it does do is, it changes their terms of service?
Trent: Exactly. So overall, like I see that it’s a distribution, right. But right now, you know, if people on Facebook, they can extract your data right now, thanks to the GDPR. And it’s actually even nicely machine readable as part of GDPR rules. But there aren’t, there haven’t so far have been compelling alternatives that have on not only built the technology, but solve the empty network problem, right. And so Google and Facebook could be in the future, holding that data, but in only in a delegated sense, right, they’ve got some access control. But the cool thing is now, US consumer can, you know, go on to your phone or whatever, and revoke certain permissions very, very easily. Right? And without and potentially delegate permissions elsewhere. So it’s kind of going be up to the consumer. But at the end of the day, it’s all yeah, it’s a choice. Overall, right now, the distribution is very, very biased towards just a small number of players, but I’m hopeful that it will be much more where there’s going to be a lot more providers for various services, whether it’s Facebook style social media, or Twitter style social media or whatever, right. And then people can loop in and out of networks, a bit more at will. And you know, if Twitter shadow bans you, for example, like happened to Ryan Sarkis, recently, you’ve got alternatives, rather than just kind of like, you know, having to deal with the craziness.
Jamie: Right. So we’re probably three quarters away through now. And the danger with me talking to you is that we could go on for two hours, and probably not even notice. So what I wanted to do was to make sure we cover off a few things. So one thing that’s really important that I want to try to explore on this podcast specifically is the kind of barriers towards Web 3 now clearly, for you, as for me, you know, data is data sovereignty, and a new data economy is central to Web 3. You know, you’ve built this protocol built this tooling. And you know, I’ve seen now you’ve kind of got these kind of hackathons competitions going, I know you’re exploring how grants could be deployed. What kind of middleware do you think needs to be built on top alongside Ocean for the full vision to happen? And what applications are you seeing most, what most exciting applications using being built on Ocean today or you would like to see being built in the short term mid term with what’s possible?
Trent: Right. So maybe I’ll start with just some exciting applications. We had this Data economy challenge that we ran from September until January. And we had submissions from, I think it was more than 25 really great submissions. And of that we gave prizes to the top nine. And they were a lot of varieties of data marketplaces, ones that were you know, pop ups in the browser, data wallets, in various shapes and forms supporting streaming data better. Supporting very specific use cases, such as molecule data, for example, in pharma. So those are all quite exciting. And we’re, you know, our ecosystem team is proceeding with sort of more of those with hackathons, and so on. So that, so that’s very exciting. Because, you know, the community can teach us and teach each other about what’s important and so on. And like I said, before, you know, we’re not claiming to know everything about what this is going to look like. So what we want to make sure is that we reduce the friction for people to build things of value, ultimately, for the overall, you know, open data economy to be successful, or the Ocean version of that, which, you know, hoping will be the one or at least part of it. People need to make money to be self sustaining, right? They need to be able to feed themselves. And the wonderful thing about making money is that if you’re doing well, you’re making, you’re taking in more money, then you’re spending, then you grow, and it’s a positive feedback loop. It’s, you know, the snowball gets bigger and bigger and bigger. And this is, you know, capitalism of the past. But we have to reconcile this now with the world of Web 3, right. So we still need to think about how do we make the overall ecosystem sustainable. And for us, part of this is, you know, helping to have seed funding for projects that are good initially, this is where things like hackathon come in, but then later on, to make it easy for people to build companies that are doing are doing things like data marketplaces, data wallets, and so on, in a way they can make money with as low cost as possible, right. So that’s kind of, you know, in terms of your question of what do we see as the essential components. essential components is, a rich set of data marketplaces for various verticals. And you know, we’re doing a lot of work on that to help reduce friction for people to roll out their own data marketplaces. I’ve mentioned already Dex rate is one of the lead ones working on in the logistics vertical. There are others that, you know, we’ll be talking about more as time goes on. Better search and discovery of datasets across you And workplaces. So there’s going to be more and more work on that. Better data wallets. So this includes leveraging existing crypto wallets that can hold data tokens directly, but then also wallets that are dedicated to holding datasets themselves. And then not just holding datasets, but then also publishing datasets, minting data tokens from that, etc.
Jamie: That’s the important bit for you around introducing data tokens because all of a sudden, you can then leverage all this existing infrastructure, Dex centralised exchanges or wallets because you’ve you’ve turned data into ERC20.
Trent: Exactly, exactly. So like Dex is, especially like automated maker type Dex, like Yoona, swap or balancer. They’re really interesting to us because they can handle the long tail of tokens, right? There’s always a buyer for a given data token then right. So you have you know, we can have a million data sets on there and every single one, it could be on unit swap for example, right. And then You can be aggregating these as well with things like your C 998. And, you know, people can be selling in where they lump a bunch of these things together into a single NFT and sell in an open sea. All these things and a lot of the defy stuff too, right? You know, we’re going to under collateralized defy. But if you’re going to do under collateralized defy, you need to be able to predict the chance of someone defaulting. Right. So how do you do that? Well, you need data. So this is going to be super, super useful. Yeah, and as well as we’re starting to see smart contract insurance with the likes of Netflix mutual. So once again, insurance has traditionally been an industry that is very data intensive. In fact, you know, that someone doing data science and insurance, it’s they, they’re called actuaries. And that label has been around for a long, long, long time. So the world of defy it has sort of two levels right. In terms of the relation to data tokens. The first is, simply you know, taking everything that exists in defy and having a data token version of it, right. Whether it’s data exchanges, you know, whether it’s data custody, whether it’s data loans, database, stable coins, all these, but the second is what I was mentioning where it’s starting to use data directly itself. And then you get really interesting feedback loops, right? So things like under collateralized defy products and insurance, right. So that, to me is very, very exciting. And, you know, this is maybe one of the killer apps will pop out as well, we will see. But this is, yeah, a big motivator for us for data tokens. And not only makes things simpler for building on Ocean, but all this you know, great potential with defy.
Jamie: Yeah. So that’s interesting, and maybe actually is a nice Segway into the last topic we can cover off in the final five minutes and he wanted to talk about, well felt, it’s impossible to not talk about what’s going on with Corona right now. So then, if you look at, I mean, obviously, the way that we talk about crypto now primarily is in the context of defy it. And you could argue that defy has been a continuum since Bitcoin it’s just that we call it defy now, but effectively, we’ve been building a bottom up capital market, proto capital market. And, you know, the folklore is that that was in direct response or reaction to what happened in 2008. Now, it’s interesting. You mentioned insurance. Clearly, if you look at what’s happening in the world right now, it’s not unfeasible that in the way that banks were bailed out in 2008. Insurers around the world are going to have to be bailed out. And if you think about all the liability that they have in the system, or the claims that they’re going to be having, come in for force majeure. So do you think that, that is the possibility this kind of collapse of the existing insurance industry could be the thing that creates this requirement for a decentralised instance and decentralised data to augment insurance as an extension of defy.
Trent: I think there’s a few questions in there, that are kind of coupled. So in general, you know, how much is the world going to be changing? And there’s different hypotheses and what this could be some, you know, some people, you know, some, some opinions are that it’s maybe at the end, you know, three months from now or 18 months from now, maybe it’s only changed by 20%. Other people are saying that everything is different for good. And, you know, what, what is different? is also of great different opinions. So, you know, with the question around insurance, then, will we see, you know, a crumbling of traditional insurance and a rise of, of decentralised? I think it will happen over time. Will it happen in the next 18 months? I don’t know. You know, overall, I do see that the the events now can be catalysing usage of, that said, you know, Andreas Antonopoulos was on the defiance podcast about four months ago. And it was a really great episode and the host of defiance Peter, he had asked Andreas so like, what is your thoughts about Bitcoin and other cryptocurrencies sort of taking over the traditional fiat currencies sooner rather than later? And Andreas raised something I hadn’t heard before, but it really stuck with me. And he said, you know, well, first of all, no one wants to wish the economy to collapse just so that crypto we better. Andreas stated that. I see that too, obviously. Right. But the second some people do that. Yeah. The second thing is you have the crypto has to be big enough and mature enough to be able to handle it. And Andreas was pretty confident that it wasn’t big enough and mature enough yet. And I’m leaning towards agreeing with that. If you think about gold, it’s a $4 trillion asset class, Bitcoin it’s on the order of 100 billion. So it’s a 40 x difference, right? So can crypto, you know, is the lifeblood of crypto big enough yet, and it’s something we shouldn’t necessarily hope that it is. I think it’s better to say we can help, you know, and the people who want to use it will. But we shouldn’t expect that, you know, the whole shipload of people is going to jump into the lifeboat. We shouldn’t be pushing too hard, because that that might be too much for crypto as well. We don’t know. Right? So, but we should definitely be trying to help however we can with this crisis, of course, and we’re taking steps ourselves that way. But overall, you know, we shouldn’t be hoping for disasters, just so that crypto can help whether it’s a more broad thing like I was talking about, or for insurance, specifically, right.
Jamie: And clearly this links to, I mean, many people might not be aware that you really catalysed the token engineering community. And you’ve always insisted upon thinking about tokenized systems, from an engineering perspective in that, to kind of speak back to the point you made about being able to bear the load as a system. The economic load in this context, we probably don’t have time to go into much detail on that now. But what I would highly recommend is that people search out the token engineering community, we’ve supported it in a couple of cities. But it is, it’s one hell of a movement now. And if you go along to one of these meet-ups, that kind of the diversity of backgrounds and disciplines that are kind of being brought about into this space is quite mind blowing. So then if you want to say, a final point on what’s going on in that community and how people can get involved.
Trent: Sure and also, if it’s all right, I’d like to just bring up after that another point. But on token engineering, yeah, I found as we were designing the initial ocean token that we were flailing, frankly. And so, we thought about methodologies that we had used elsewhere and realise that the methodology we’ve been using for, I used for design of optimization systems, you know, and other AI systems had worked very well. And it also looks a lot like just other more traditional things, you know, define the problem well, try existing patterns and, and, or try to change the problem definition if you can, and only if needed, you come up with something new, because novelty is risk, right? And, you know, if you’re deploying, you know, hundred million dollar chains, billion dollar chains, then you want to minimise the risk, not maximise, right? So keep the things simple and small. So that’s a, that was sort of the methodology we came up with. But also, you know, the reason I chose the term token engineering was on purpose. Engineering implies a very specific way of thinking. I’m trained as an engineer. You know, we had classes on ethics and all this. And the first thing we saw in the ethics class, as well as it was 10 times before that in engineering was, you know, the collapse of the Tacoma Narrows Bridge in Seattle. And it’s because the engineers hadn’t been thorough enough to account for oscillations caused by resonance from the wind. So and, you know, that’s, it feels like an esoteric thing, but really, like, it was an overpass that had a lot of wind, so you should be accommodating. And so, you know, one way to think about token engineering is that it is a combination of theory, practice, tools and responsibility for designing, deploying, and maintaining engineering systems, right? And so the responsibility thing ties into the ethics part. And that’s something that had been, I think, under looked under emphasised or overlooked and underrated. Often in the crypto community, people had been talking about it here and there. And I realised, hey, you know, like this has been, this is a staple of engineering. It always has been, you know, in Canada, if you want to be an engineer designing bridges, you have to get accredited by a particular institution in Canada. And so why not bring a lot of thinking and learning and long standing practice into the world of designing this new infrastructure for civilization, which is blockchain technology? Public blockchains.
Jamie: Yeah, I mean, what, what a great way to end I love that idea about introducing Responsibility and Ethics into designing in depth into designing these systems. So I’d hoped to do 45 minutes we’ve gone over 50 there’s so many more things that we could talk about. I’ll have to get you on again, Trent. Thanks for coming on the show. Really good to chat to you. And I’ll share the links both to you, Ocean token engineering, in the comments.
Trent: Thank you for having me. Appreciate it. Awesome.
Jamie: If you enjoyed today’s podcast, please make sure you subscribe rate and share your feedback to help us reach as many people as possible with the important mission of Web 3.