By nature of his background as a mathematician by trade, as well as his history as one of the co-founders of Ethereum, Cardano founder Charles Hoskinson is uniquely positioned to understand the importance and interplay between theory and application when it comes to blockchain development. Finding the right balance between these two disciplines is very difficult, yet critical to achieving mainstream adoption of crypto. Whichever blockchain solves this puzzle first will likely become the dominant platform in the years ahead.
That said, these topics can seem overwhelming to the crypto curious, as well as enthusiasts that have been involved in the industry for some time. Therefore, I connected with Hoskinson to have a substantive discussion on the theoretical underpinnings of crypto, consensus mechanisms, and distributed computing and explore how their evolution falls within the canonical history of mathematics and computer science. This interview is a bit technical, but he does a terrific job of breaking down complex topics into simple terms, often through the use of examples.
Forbes: You’ve made a concerted effort to include academic and scientific rigor into the development of Cardano and crypto as a whole. Can you please explain the process?
Charles Hoskinson: We’ve published 102 papers over a three year period. A big part of those 102 papers was creating strong theoretical foundations for cryptocurrencies as a whole, not just for Cardano. For instance, we wrote a paper called GKL15, which has been cited more than 1,000 times, and it’s the canonical way of looking at what is blockchain.
Another part of the portfolio is industrial research, where we said, “Okay, now that we have the theory, what can you do? Can you shard (partition) proof of work? Can you shard proof of stake?” Let’s go build a proof of stake engine. And that’s not necessarily protocol-specific but it gives you a sense of capabilities. For example, if I want to have anonymous medical records or if I want to build a global scale system with billions of settlements. What do you need to actually do with a protocol designed to make that happen?
And then the third part of our research portfolio is protocol-specific. How do we take those capabilities and actually put them into a system? Now, we go to engineers, a very special type of engineer called “the formal methods engineer”—they read a scientific paper and actually create the blueprint. It’s almost like an architect to a general contractor; the architect draws these blueprints up, shows you how to make the house, but they obviously don’t build it—the general contractor does that. That’s the hardest part.
Forbes: Many people coming to crypto and blockchain may not realize that the ideas of decentralized consensus and networks didn’t originate with blockchains. Talk a little bit about how that work, with roots in computer science, informs the research you’re doing as it pertains to blockchains?
Hoskinson: Distributed systems is one of the oldest areas of computer science—conceptually a very simple problem but, in practice, a very difficult one. One of the pioneers was Leslie Lamport. He wrote some of the foundational papers—for example, Lamport clocks, about how to keep time in a distributed system. That was back in the 1970s. He also wrote Paxos, which is a distributed systems protocol. That was the first one that was called “Byzantine resistance.” But the basic idea is that the minute you leave the comfort of your laptop or your cell phone or your computer and you go into the web—a distributed system—then your perception of events and reality is different than other people’s perception of events and reality.
For example, let’s say you’re closer to Michael’s computer than Jenny’s. If something happens on Michael’s computer, you think that happened first, but Jenny would think that happened second. So when you order these things, Jenny would put: Michael – 2. You would put: Michael – 1. The point of consensus algorithms and timekeeping is to create one logical clock or one logical, canonical ordering of events. Why is this important? Let’s say you’re running a financial system. Who got the trade if there are two bids that come in at the same time? Did it go to Alice or Bob? Well, based on how you order things will determine it. So, it’s really easy to say: “We’ll just pick a common point of reference—a central server, like a Microsoft server in Washington, and whatever ordering it gives us, that’s great.” But when you are actually in a distributed system where no one is in control or special over anybody else, it turns out that’s a much, much, much harder problem. Especially if you admit what are called Byzantine actors—people who could lie and cheat.
We’re definitely a big contributor in this space, but there are a lot of great teams like Algorand, for example, and the Cornell team led by Emin Gun Sirer, the Snow White protocol, and so forth.
Forbes: This leads me to the proof of work/proof of stake question. One of the signature features of Cardano was the Ouroboros consensus mechanism. Many may not know that there are different variants of proof of stake. Can you break it down?
Hoskinson: First off, whether you’re proof of stake or proof of work, you can have three things that you have to accomplish with a consensus algorithm. First, you have to pick somebody to be in charge for a moment of time. And that could be a block or an epoch, but some unit of time. And that person has to do something with that power, so they make a block. Then they have to broadcast that to everybody in the network, and the network has to accept it. It’s kind of like a poker game: you have to pick somebody to be the dealer, and that person shuffles the cards and deals them. Then the players pick up the cards, look at the decks and either choose to accept them as they are or reject them. So for example, if you get a hand of cards that has five aces you would say there’s something fundamentally wrong with the deck. Because the deck should only have four aces, right? A consensus protocol does the exact same thing.
The primary difference between proof of work and proof of stake is that first stage—the picking of somebody to be in charge. With a proof of work system, it’s a meritocratic mining process; basically, you’re mining hashes, and those are like lottery tickets. You just keep going until you hit the lucky numbers. Then you say, “I have a golden ticket.” And that gives you the right to make a block. The higher the value of the asset, the more competition you get, the more energy expenditure. 99.9999997%, if not more, of the energy consumed by bitcoin is from that first stage. The other two is everything else.
With proof of stake systems, instead of mining we say, “We’re just going to go ahead and weight your stake proportionally, treat it like a synthetic lottery, and you should win on average that much.” The advantage of proof of stake is that because you don’t have that gargantuan overhead and energy expenditure for deciding who gets to make a block, it means you can put a lot of your magic in the other two stages. So you end up getting protocols that are much lighter and massively more energy efficient. Cardano, for example, is 1.6 million times more energy efficient at the moment than bitcoin.
Forbes: More than 70% of Cardano is staked right now, for Polkadot it’s about 64%, Ethereum is far below. What do you think is a healthy amount of an asset to be staked during some of the early stages as well as at maturity?
Hoskinson: This is one of those apples and oranges things. When you talk about staking in a bonded system (which have lockup periods that can span days to over a year), you’re actually making a very significant statement. Because what’s happening is you’re saying your stake is locked, and people can’t move it, they can’t spend it. That means that those tokens have been taken out of the supply. When you look at staking in Cardano, they’re always liquid, there’s no bonding required. So that means that’s still a liquid stake, and that’s one of the biggest confusion points we tend to see. With Cardano, they tend to think that means the circulating supply is much worse than it actually is.
The other thing is that our system also has voting on-chain, and outside of Tezos we’re probably the most advanced there. That means when you have your stake, not only can you delegate it, you can also use it to vote on funding proposals and changes to the chain as a whole.
Forbes: You’ve talked a lot about the importance of interoperability when it comes to other blockchains—the sense that blockchains can’t be self-contained, they have to be able to know what’s going on in the world around them, bring in outside data, etc. What’s the right way to find that balance, so that blockchains are still decentralized at core but also useful beyond some of these self-contained use cases like a payment system?
Hoskinson: What you do is focus on the ability to move information, value and identity between chains, and then you kind of let the markets decide where things are going to live. It’s funny, everybody wants to be open source until they don’t. Like Ethereum says, “Hey, we’re open source, we’re an open ecosystem. But by the way, we want to be like Microsoft, with Internet Explorer and ActiveX, lock everybody into our ecosystem!” Shouldn’t users be liquid? Shouldn’t information value be liquid? So what we focus on is our cross-chain communication protocols. We wrote a lot of papers basically describing what you can and can’t do in that context and also protocols to enable you to move value, represent value information between systems.
I think over the next three to five years what will happen is our industry will converge to a Wi-Fi moment where it just works. And no matter what ecosystem you happen to be in, it will be very easy for you to migrate from that system to the next system with some notion of a timeout. You click a button, and now you’re in Ethereum, and it takes a few minutes or a few hours until you’re there. You click a button, now you’re in Cardano. It takes a few minutes, a few hours and you’re there. That means it’s going to be a race to a bottom, in terms of operating cost. So if there’s an app that is very expensive to run in one domain, people will simply flee that domain and go to a much cheaper blockchain to operate—that’s the internet of blockchains that we’re probably facing.
Forbes: I want to finish by asking you what’s next for Cardano. What’s coming up on your roadmap?
Hoskinson: What we’ve been doing is pulling all the pieces together. We’re kind of finishing up the Cardano 2020 research agenda, and the idea is that over time as the research crystallizes, we’ll turn it to the best commercial product we can.
Ideally, you pick up a number. Ours was “let’s do it in five years or something; let’s have it all done by 2020.” Of course, I’m a very optimistic person and I tend to underestimate engineering and scientific complexity, so we didn’t quite hit that milestone. 2021 is kind of our overflow year, where we’re pulling together and turning on all the things we dreamed of over the last five years. That includes things like our governance stack, smart contracts, our metadata standard, token issuance, full decentralization. We’ve hit almost all our goals: we’ve done metadata, we now have a native token. There are more than 10,000 tokens issued on Cardano in just a month, which is mind blowing. And in the next few months, we are turning on smart contracts. The last mile from there—we will be turning on all the governance components. We already have 20,000 people participating with them, but they’re running as a side chain, so some of those things need to be linked to the main network. Once those things are turned on, commercialization of the platform is already underway. We’re getting nation states to do cool things and we’re bringing millions of users in through that push model. We’re also doing decentralized finance, non-fungible token marketplaces—all the usual suspects that you see in the Ethereum lab.
Forbes: Thank you.
