IEEE Blockchain Podcast Series: Episode 7

A Conversation with JP Vergne
Associate Professor at University College London, School of Management

Part of the IEEE Blockchain Podcast Series

Episode Transcript:

Brian Walker: Welcome to the IEEE Blockchain podcast series, an IEEE Digital Studio Production. This podcast series entitled Research Notes in Blockchain is hosted by Quinn Dupont, former assistant professor at the University College Dublin School of Business, and founder of Alumni, a Web3 startup with a mission of putting university diplomas on blockchain. Quinn is also the author of Cryptocurrencies and Blockchains. In this episode, Jean-Philippe Vergne, associate professor at University College London School of Management discusses his research on how blockchain technology can be employed to change how organizations are structured at the managerial level. He also explores blockchain as an underlying technology in the development of a new digital platform for creating end user value, as well as providing insights on the differentiation of decentralization and distribution.

Quinn Dupont: Thanks, JP, for joining me here. One of the things I want to first talk with you a little bit is actually your transition from Western to University College London, and I knew you back when you were at Western, and you were doing really interesting work with the Scotiabank FinTech stuff there, and but I've since seen that there's been a lot of action at UCL, and so I was wondering if you could maybe just give me a little bit of sense of what's happening there in blockchain research, and what are you really working on right up to the minute? Like what's your next kind of research project? If you could give me a little sense on that.

JP Vergne: Sure. Thanks for having me. I'm very happy to chat with you today. So the UCL Center for Blockchain Technologies is a transdisciplinary initiative within University College London that bridges people from computer science, engineering, economics, and more recently, business to create a platform for research on blockchain, broadly speaking, and there's also industry associates and industry partners, including some large blockchain ecosystems that are all kind of hanging out together and sharing ideas and data to advance research in the space, and my specific angle these days, I would say is twofold. The first one is, as a researcher in organizational sciences and management, I'm especially interested in how blockchain can be seen as an organizing technology that can create new ways to design organizations, and if I had to like use a bit of a formula to summarize what I'm interested in, that would be the idea of having management without managers. So having the possibility to design organizations without having managerial authority, and I think blockchain is particularly promising in this respect. After all, we now have large organizations such as Bitcoin or Ethereum that are operating mostly without managers, and they've become global. They've become large and they've grown quite successfully. So, there's a new model here, and I'm really interested in how it can be applied beyond cryptocurrency, and then the second area of research that I'm looking at is basically looking at blockchain as an underlying technology for a new type of digital platform. A lot of the large digital platforms that we have out there, whether it's the ones run by Facebook, or Google, or Tencent, or Yandex in Russia, they have in common the fact that they rely on machine learning as their core technology to create value. I think blockchain can provide an alternative template whereby we can have decentralized platforms that create value in different ways, and that's what I'm looking at these days as well.

Quinn Dupont: That's fantastic and so that's perfect. That leads right into this paper that you published last year in organization theory called Decentralization-- sorry, “Decentralized vs. Distributed Organization: Blockchain Machine Learning and the Future of the Digital Platform”, and I wanted to talk to you about this paper specifically because, well I mean, to be honest, it's one of the more interesting papers I've read in some time where you compare a couple of different dimensions of blockchain and machine learning and offer some insights into where these might go, and specifically the kind of organizational forms that they cultivate. So maybe you could just say a little bit more about this distinction between decentralized and distributed as you put it in the paper.

JP Vergne: Yeah, sure. So, if you look at the space today, whether it's on the academic side or the industry side, people are often talking about decentralization, and how decentralized they are, and whatnot, and what I've been wondering for a while is what do they actually mean by that? It was for a while a PR tactic with unclear technological underpinnings. It's become a more important issue recently because it really affects regulation and public policy. So, for instance, there's been claims made by officials of the Securities Exchange Commission in the US saying that if your network that is decentralized enough or sufficiently decentralized, then maybe the tokens that you issue may not be considered securities, and so they may not be regulated as securities. So, the question becomes, okay, what is decentralization really and how do we measure it? So, looking at this particular issue, I looked at what people have been saying about this topic in the field of computer science, in the field of management, economics, network engineering, and you have very, very inconsistent definitions that are provided about decentralization, and people really mean different things when they talk about this. It's a very vague notion. So, looking at extent discourse, I actually went back to the fundamentals of organization theory and basically, emphasized the idea that when you have an organization there are two important dimensions that you need to take into account to be able to achieve anything, and the first one is how you process information, and the second one is how you arrive at decisions. So, information and decision making, and then I noticed that the two dimensions, they're correlated in some way, so to make decisions, you need information of course. However, they don't have to overlap. So you can have information within a particular organization, or within a particular social network, or within a particular digital platform that is more or less dispersed across actors, across members, across nodes, and similarly, you can have decision-making authority that is more or less dispersed across members, across nodes, across agents, and so looking at these two dimensions separately, it is possible to have more dispersion on one dimension, less dispersion on the other dimension, and out of convenience, but also to be in line with prior thinking on the issue, I have proposed to define decentralization as the dispersion of information, and distribution as the dispersion of decision making. So, you can be more or less decentralized on the one dimension, and you can be more or less distributed on the other dimension, and basically, this distinction helps us understand a number of phenomena that have been out there without really an explicit conceptualization. For instance, to pick a very practical example, the idea that we can have distributed ledgers that are not decentralized, that are actually operated by a corporate entity in a very centralized fashion, and that can really help us understand what is the difference between public permissionless blockchains such as Bitcoins or Ethereums, and on the other hand, corporate IT projects that are implemented by large corporations to cut costs in their supply chain, for instance, and in the latter case, they'd be using distributed ledgers, but they're not necessarily decentralized.

Quinn Dupont: Yeah. I think this is one of those really interesting fundamental distinctions, and what I like in your paper is that you then apply this sort of-- you come up with a fourfold matrix to talk a little bit more specifically about different kinds of organizations. So you've got, I guess it's centralized and concentrated. You've got decentralized and concentrated, centralized and distributed, and decentralized and distributed and they each have their own kind of characteristics. Can you say a little bit more about sort of that fourfold matrix?

JP Vergne: Yeah, so when you look at these two dimensions, you basically can describe four ideal types in a way, where you can be centralized and concentrated on one extreme, and then you can be decentralized and distributed on the other extreme with two intermediate scenarios, and so a lot of traditional corporations, they use delegation of decision making to middle managers, and so there is, for instance, distribution of decision making among these organizations. However, there is still centralization of information along that managerial backbone, and so for instance, when we look at some of the recent controversies involving Facebook, we could see that, despite claims to the contrary, Facebook was still operating in a very centralized fashion, for instance, when it was all the way up to the CEO to decide whether a particular account on facebook.com during the presidential elections should be banned, or authorized, or whether a particular post on facebook.com should be deleted or flagged with a warning. So, imagine that in a cooperation of like more than 10,000 employees, and billions of accounts, and several billions of posts made on a weekly basis, that it's the number one CEO person who still has the ultimate information, access to the database that records posts, and status updates, and things like that, an account, right? So, this is a clear indication that Facebook is distributed because it relies on managerial delegation, but at the same time, still very centralized. Now blockchain-based organizing can enable the simultaneous distribution of decision making and decentralization of information processing, and I think this is the true novelty of blockchain. We are still in the early days in terms of understanding what the implications of that are. But it's an incredibly promising innovation. I would place it on par in terms of how radical it is, with the invention of the publicly traded corporation or the limited liability corporation. I think this is when we will look back at history, blockchain-based organizations will be up there with these other radical innovations.

Quinn Dupont: That's yeah, I agree. I think that's really a fundamental point to make. Would you put something like a decentralized autonomous organization, a DAO on the kind of polar opposite as far as the spectrum goes, like this idea that using DAOs we can decentralize, or I should say distributed I think, in your nomenclature, decision making?

JP Vergne: Yeah, I think they are probably actually the perfect example of that. Provided that every DAO member has access to all the information, and there's no restricted access for particular members, then we'd have full decentralization in terms of information, and provided that every DAO member that owns some of the governance tokens has a right to vote on all decisions. So, there's not like a multilayered hierarchy of decisions where like minor decisions you can vote on, but the very important decisions you're excluded from voting on them. If you have this full access of members to decision making and all sorts of decision making as provided that they have governance tokens, then you would have full distribution as well, and so yeah, DAOs clearly have the potential to enable that. Not all of them so far actually are working in this way, so we have to be careful. But this is typically something we don't see in traditional corporations. Now the fact that it's theoretically possible and actually practically happening already does not mean that it's efficient or more efficient than alternative designs. So, it is very experimental at this stage. It may be more democratic, but it will not necessarily lead to more successful organizations, or how successful they will be might depend on the context and what they're trying to achieve. Because the reason why we have traditional corporations the way they are designed where some people don't have access to all the information, and some people don't have a say in every decision-making process is basically because of the division of labor, and so there are reasons why it may be more efficient not to be decentralized and, or not to be distributed.

Quinn Dupont: Yeah, and then this is one of-- so this goes back to that point about the efficiency that comes from your standard, sort of your traditional organization that has a managerial layer.

JP Vergne: Absolutely.

Quinn Dupont: So, on that point, let me switch a little to kind of the second part of your paper, which talks about this really interesting tension between really two hype kind of technologies, blockchain and machine learning, and you suggest that really, there's a fundamental tension in terms of the organizational forms that emerge out of these two technologies. Could you say a little more about that?

JP Vergne: Yeah, I think that the tension becomes visible when you compare the business models of centralized digital platforms like facebook.com or Instagram and the operations of decentralized and distributed platforms such as Ethereum or Bitcoin. You really see that when you are relying on machine learning, in the former case, you really have strong incentives because of the nature of the technology to pool together a massive amount of data, so you can derive predictions that you will then monetize one way or another, and this is what Facebook is doing. They are monetizing behavioral predictions about their users, and they're selling them to advertisers, and to be able to do that you need a critical mass of data, and there's this pool. So, we can speak of data gravity. There's this mass of data that becomes accumulated, and the more data you have, the more data you attract because you're going to attract third-party developers and complementors to your platform, and more data will lead to more predictions, which by themselves become data as well, and et cetera, et cetera, and so on, and so on, and so you can see how a centralized form of organizing is actually a pretty good fit with the use of machine learning as a core technology to create value. Now on the other hand, if you want to achieve value by having transparency, maintaining an independently auditable track record of prior transactions and ownership patterns, it's very important to guarantee that any user can publicly access these records, and this is how you can maintain trust in the record itself, and so because of that, you will want to decentralize information as much as you can, and so a good way to do that is to use blockchain, and so it's a very different way of creating value. Obviously, it's for different purposes as well. But so, I would say if I had to summarize the implications of these two technologies in terms of how they produce trust, I would say that machine learning is producing trust in a distributed fashion through the use of corporate hierarchies, and blockchain is producing decentralized trust using blockchain in communities that are flat.

Quinn Dupont: Yeah, this issue of trust seems to underpin so much of blockchain. It's interesting to see it coming up in organizational forms here.

JP Vergne: Yes, I think it's a fundamental issue, and I think that when we talk about the next big thing for the internet, which is called by some the Metaverse or the multiverse by others, I like another term myself, which is the paraverse. But anyway, that idea that we're going to be spending most of our time potentially navigating across virtual worlds. The question is, how do we connect those different worlds and how do we ensure that whatever value is created digitally in one of those worlds can be used and exchanged in a neighboring virtual world, and so we're going to need to establish trust to be able to create value across virtual worlds. So, there are two ways to do that. One way is the corporate way, which is what Facebook now renamed Meta is proposing, which is to create a corporate umbrella that will create standards, and that will have corporate control over a number of virtual worlds and platforms. Some of them will be social networking. Others will be gaming, and you'll navigate them using maybe a piece of hardware that's owned by the corporation. Perhaps it's going to be a VR headset, and that will guarantee whatever reputation you accumulate in one world is recognized in another one, and so you have to trust the managers of Meta that they will enforce these mechanisms to ensure that value is maintained over time, and so that's distributed trust because you have to trust managers to whom decision making has been delegated. Now on the other hand, if you look at some of the blockchain ecosystems that are created out there, whether it's Decentraland, whether it's to a lesser extent, but still Axie Infinity, or what is going on today in the Polkadot ecosystem with its parachains, you have a very different way of maintaining trust. You don't necessarily have a corporation in those cases, and the value that it’s creating in a particular virtual world is maintained by a decentralized record, and these various decentralized records, we can call them blockchains, or we can call them parachains, they are interconnected with each other using open-source protocols that make them interoperable, and so instead of having a corporate umbrella that kind of oversees everything and can capture value everywhere, instead you have an archipelago of islands that are loosely connected using blockchain technology, and so in that case, I prefer to speak of the paraverse in reference to the parachains that Polkadot is incurring to its main chain in order to maintain trust, and so I think these are two very different models, and I feel that soon enough they will be competing with each other. It will not just be Facebook competing with Microsoft and with Google in the Metaverse. It will be centralized platforms competing with decentralized platforms, and that's a very exciting battle to be had. I don't know where it's going to go but I think it's just starting.

Quinn Dupont: Yeah. I mean that battle, I mean you mentioned here Facebook, and Meta, and the Metaverse or paraverse, as you might want to call it. This gets us right to the question of regulation, which is something you pick up in this paper as well, and you propose this really interesting idea of kind of a data-level regulation as a way that we can maybe break some of these existing concentrations of power. Can you say a little bit more about what you mean by this sort of data-level regulation?

JP Vergne: Yes. So the traditional approach to regulation has been to focus on the corporate level, and so you take a corporation, and you say, “Okay, how much pricing power does it have? Does it have too much pricing power? Does it have a market share that's too big?” I mean if you answer yes to those questions, then maybe you consider breaking up that corporation, right, forcing it to divest some of its businesses, and really, that works well when you are looking at centralized structures, and when you are looking at traditional markets where the product is not just data, and when you look at dominant digital platforms today, really, the business they’re in is the business of data, and it doesn't make a lot of sense to break up corporations if they are allowed to just part ways having each a copy of the original datasets, for instance, because then you haven't achieved much in terms of weakening their dominant position. You have actually potentially created a duopoly instead of a monopoly, or you have actually helped them specialize each on a different niche market, and so when data is the core material that generates value from which you create value, I think it's important to realize that this is where regulation should be happening, and so the questions that the regulators and policymakers need to answer these days are, okay, how are you allowed to collect data? When are you allowed to deanonymize data? When are you allowed to sell predictions on this data? When are you allowed to prevent users from taking their data back or porting their data to a different ecosystem? How do you make sure that when that happens, the original record has been deleted, things like that, and this is really the condition under which we will have fair and vibrant competition in the future of digital platforms.

Quinn Dupont: Yeah, I think that that hits the nail on the head there, fair and transparent competition. So with that, I think that's probably a great place to end. Thank you, JP. It was much appreciated. I really enjoyed talking to you today.

JP Vergne: Thank you, Quinn. It was a pleasure.

Brian Walker: Thank you for listening to our interview with JP Vergne. To learn more about the IEEE Blockchain Initiative, please visit our web portal at blockchain.ieee.org.