🎙 Polkadot's Gavin Wood on Building a Layer 0 to Underpin All Blockchains
Gavin Wood’s crypto journey started with Ethereum – he was one of the eight original founders and the project’s CTO in the early days. Today though, he’s busy building Polkadot, a chain that is attempting to improve where Ethereum falls short. After years ...
Gavin Wood’s crypto journey started with Ethereum – he was one of the eight original founders and the project’s CTO in the early days. Today though, he’s busy building Polkadot, a chain that is attempting to improve where Ethereum falls short. After years of development, the ecosystem’s first applications are finally coming to life as it lays the foundations to become a layer zero, where multiple Layer 1 chains and their decentralized applications can flourish. Polkadot aims to give developers the flexibility to build decentralized applications as they see fit, and on scalable infrastructure.
Gavin wants to turn Polkadot into a multichain powerhouse where financial systems and web apps can run and interact seamlessly. To Gavin, Polkadot can be the base infrastructure for a web3 future, a set of technologies that can guarantee privacy, trustlessness and decentralization, which to him, will be crucial if the free world is going to survive.
The podcast was led by Camila Russo, and edited by Alp Gasimov and Daniel Flynn. Transcript was edited by Samuel Haig.
🎙Listen to the interview in this week’s podcast episode here:
🙏 Thanking our podcast sponsors:
- Zerionis Mission Control for Web3: an intuitive DeFi portfolio manager, multichain tracking & trading and the best place to show off your NFT collection
- Unstoppable Domains is The #1 provider of NFT domains. Get YourName dot crypto, dot x, dot nft or a range of other endings for as low as $5
- Nexo: Invest in crypto, earn interest of up to 20%, paid out daily, and use crypto as collateral to receive a credit line at premium rates.
- Step Finance's portfolio management dashboard enables Solana users to visualize, analyze, execute and aggregate transactions across all Solana contracts at the click of a button.
👀 Only paid subscribers have access to the full interview transcript below.
Cami Russo: All right, here we are with Gavin Wood. Gavin, I'm so thrilled to have you on The Defiant podcast, thank you so much for joining me.
Gavin Wood: Lovely to be here, Camila.
CR: Gavin is the founder of Parity Technology and the Polkadot blockchain. He is also co-founder of Ethereum Polkadot is now the eleventh largest cryptocurrency by market cap at around $20B, while the chain is the sixth largest Proof of Stake chain with around $11B in staked value. The last few months have been pivotal for Polkadot with the parachain auctions going live, laying the groundwork for applications to be built on the Polkadot ecosystem, and we'll get into all of that.
But before we do I wanted to start with your pre-Polkadot life and how your background and your path in crypto started — obviously with ethereum So can we start with how you got interested in crypto, and your role at Ethereum?
GW: Sure, well I got interested in crypto, I think I saw a story about Bitcoin back in 2011 on the geek news site, slashdot, and I thought about buying some then, but then it was too much effort so I didn't bother — I guess a lot of people are the same.
And then it came up again in 2013. It was a story in The Guardian newspaper about these kind of… Cypherpunks. And they had this guy, Emir Taaki, who was living in a squat at the time in south London, and it was a pretty romantic kind of article in the classic sense of romanticism. It had a picture of this hoodied guy that hadn’t changed his clothes in quite a few months with the city of London — one of the major financial centers of the world in the background — with Emir basically saying ‘we're taking on these guys’.
This came not that long after the financial crisis. It was early 2013… so it's still very much fresh in everyone's minds, and it was an interesting article — it was enough to get me to reach out to this guy and let him know I was interested in this world. I suppose it was from there that I met some of the other people in the crypto ecosystem… We didn't talk that much to Emir until about a year later when I got to know him [after] having co-founded Ethereum. Back then he introduced me to a couple of interesting people, one of whom I would later employ, and the other that I would stay in touch with for a long while.
CR: What about this idea of defying traditional finance? You mentioned this romantic idea of these techies and geeks living in a squat, wearing hoodies, looking at the city of London in the background that appealed to you — why do you think it did?
GW: I don't know if the idea was wanting, but looking at how long it had been since he'd had a wash, the thing that was romantic was more the image that it presented, which is to say there were some scruffy coders that were trying to take on this demonstrably broken financial system, and they seemed to have a pretty good idea about what they were doing.
So it was that coupled with some of the other stuff that was going on at the time. Early 2013 was basically when Silk Road came to prominence and there were a few articles that I read about that as well. I wasn't really into this area of technology… up until then, and there was a particular article I seemed to remember reading where the author of the article had actually ordered some stuff off Silk Road for journalistic reasons in order to demonstrate that this was actually possible, and apparently not too difficult. And it was this, together with the romantic imagery from the squat… It gave me the impression that this was going to be quite disruptive and it was going to cause change, and that in and of itself was enough to make one curious, and I wanted to understand the technology better. Part of understanding a technology is, for me, at least, implementing… You never really know if you truly understand an idea until you can actually use the idea.
CR: Do you think by now, is it fair to say that initial inkling that blockchain and crypto would disrupt the financial system — do you think that's already happened? Do you think that it is safe to say that you were right in thinking that back then?
GW: I think I was right in identifying that it was a disruptive technology. But I don't think that major disruption, at least to the financial system, has really happened yet. I think the main thing is that this technology is gonna keep getting developed, it's gonna keep getting polished. It's gonna keep becoming more usable.
One of two things will happen. Either. It will become usurped, and much like the internet did, it will fall prey to centralized interests. Or it will demonstrate itself to be absolutely necessary for society to move forward, and it will do this simply by there being more financial meltdowns in the traditional finance system, and it will provide the alternative path. Whether [that is] because it's actually outperforming the traditional financial services sector day-to-day, or whether it's necessary because the traditional financial services simply cannot provide the services anymore because it's broken, I don't know.
I don't know which one of those two it's going to be, but assuming that this is the disruptive technology that I identified it as, then one of those two things has to happen.
CR: Okay, so to you, blockchain tech is still proving itself, it's not there yet. What would you have to see for that to become a reality?
GW: I think there's a while to go in the journey yet. To be honest, the financial side of this has never interested me that much. Sure it's kind of interesting, but the thing that kept me interested is really this concept of an applications platform, this alternative web. And in those terms, what I would really be looking to see is if someone comes along and wants to make a web service that everyone could access on the internet.
The web has traditionally been the set of technologies and protocols that you will use to implement that, and the thing that I would look for is new developers that are coming to this and looking to create a service. Are they using Web3 technologies or are they using more traditional centralized web 2.0 technologies? And if the technologies get to the point… that newcomers naturally go to Web3 than the web 2.0, then I would say we have reached the point of adoption.
CR: Interesting. Okay, so if the default for developers building web apps becomes blockchain and Web3, then that means that blockchains have won, or they have disrupted the system, right?
Speaking of platforms for applications, how did that initial foray into investigating what crypto is lead you to Ethereum and what was your role there?
GW: It led me to Ethereum via a couple of these people that Emir Taaki met, one in particular called Johnny Bitcoin. He was a good friend of Vitalik, and it's through him that we got introduced. Then fast forward to December 2013, a few months later, and I was having a pint with Johnny down the pub in London, and he was mentioning that there was this project that Vitalik was looking to kick off called Ethereum. [Vitalik had] written this Whitepaper, and since I was such a good coder, he believed I should help. So I said ‘yeah, all right, fine, I'll do it’, and really it was curiosity. It was like I was interested in learning more about it, but I wasn't so interested in reimplementing Bitcoin. But it was an interesting proposition to implement a successor protocol.
So I took a look at the Whitepaper and I started coding it a few days later, and that led me to the rest of this motley crew that were trying to found this Ethereum protocol. And in the weeks that followed, I sort of went from core developer to CTO, basically because there was no one else that understood the protocol better. And then we got the Swiss Foundation sorted, sort of moved in some sense to Switzerland, and then all the rest of it… It was a very interesting year or two.
CR: Awesome. So you were at the core of Ethereum through its early years — the development of the yellow paper, then through the launch of the Ethereum mainnet. And then by 2016, you left the project and started EthCore, which later became Parity, developing one of the main clients of Ethereum. At what point did you decide ‘now is the time to start my own chain, and what inspired this?
GW: A few things. So from 2016 to 2017, I was still doing a lot of Ethereum coding… In 2017, I moved on a little bit into UI coding, doing some stuff around that, but it was late 2016 that I was really just thinking aloud how we might do a sharded chain — do a scalable version of Ethereum. And it was while I was thinking this through that I realized that one of the elements that would be part of this new sharded blockchain, this very scalable blockchain, was that it could be more general than I had previously assumed it would be for a sharded blockchain. And the way to make it general was going to be using this language, WebAssembly.
Now WebAssembly is like a language that allows web pages to specify computer programs that work on any browser, and they generally work fairly fast. They can work usually about 50% of the speed of a normal computer program, and this is really good. This allows the web as a platform to exist, so this is why we can have video games on the web. It's why we can have stuff like Google Office and Google Docs and all these sorts of things — it is one of these really important components that turns the web into a real applications and games platform.
So essentially the idea was ‘let's reuse this technology, but let's do it so that each of these individual shards… could have a much more specialized piece of work that they focused on. And the same way that one website might be specialized for a particular game and another website like Google Docs might be specialized for writing word documents, we will do the same but with blockchains. And the idea is that you have these specialized blockchains and you tie them all together. So they're actually just shards of a single blockchain, and that means that they can communicate with each other. And it also means that they could use the same underlying capital pool, the same underlying staking, to secure them.
And this, I felt, was a significant technological difference. It made me want to work on it because I really didn't want to work on Eth2… back in those times I would have preferred just to have continued working with Vitalik, but a couple of things transpired. One of them was [this sense of] ‘hey, here's a different idea and it's interesting to explore’, so curiosity was pulling me in a different direction. Secondly, there really wasn't that much activity going on Eth2, it was very much a research project back then, and even a little bit still now. So there wasn't that much to actually start coding and I really wanted to be writing stuff that was going to be used.
And there was another reason as well, which was that I felt that the governance processes of Ethereum were opaque and wishy-washy, vague. I didn't much like this rough consensus [where] everything happens offline — there were no clear rules. I felt that one of the really important things that blockchain had provided us was the ability to have to say ‘here are the rules and they will definitely be enforced, they'll definitely be followed and it will be transparent and… open. I felt that though [blockchains] had obviously been used for financial services with the Bitcoin currency being the first one, it would be really good to use them also for governance services — that we really needed to govern these decentralized protocols. So I was also drawn to making a protocol that had governance built-in, and… if it had governance built-in, wouldn't be great if we could tie it directly into the upgradeability so that we don't have to do these horrible hard forks that [have] oftentimes gone wrong on Ethereum, [and] take an awful lot of of of extra protocol consensus — basically a lot of human consensus, and to be honest, centralization.
If you look at Ethereum for example, it's the Ethereum foundation that specifies what hard fork is the correct path for Ethereum. So I was very interested in exploring each of these areas, and they were not things I was going to get from Ethereum. So eventually I took the decision to push forward with Polkadot.
CR: I want to dig into WebAssembly, sharding, and governance later. But first, one superficial question: Why the name?
GW: There's a few factors in the name. I like the name Polkadot because there isn't a center to the Polkadot pattern — most concepts have a center… its a centralized object… If you look at traditional iconography [and] you look at traditional naming, it’s [usually] centered around somewhere in the middle of the icon. And with Polkadot, there was this concept of it just being everywhere… In addition to that, it was kind of playful… and it was different… which I quite liked.
It was different from all of the other things that just sound like math-fi or sci-fi concepts, and I feel that… if we're really trying to take this technology to people that don't care about having the cutting edge of technology, then we need to be appealing to something beyond math-fi and sci-fi. We need to be appealing to something that is tangible, that is within people's everyday lives, so I felt Polkadot was quite nice from that branding perspective.
When… the white paper… was originally written… it wasn't called Polkadot, it was called something else. It's actually had two different names. I'm a bit embarrassed to be honest because they're very terrible names — the first one was “Interplex,” and the second was “Disparity.” I'm happy with the second but it has such negative [connotations] because we are a parity, so it's like disparity and disparate chains. But in the end, it felt a little negative, and Polkadot felt more fun and playful, and I think it's important to acknowledge that at least a good portion of life should be about fun.
CR: I love it… Okay, so Eth2 had been kind of thinking of sharding, so is it right that the main difference with Polkadot was the use of WebAssembly?
Polkadot’s model for sharding
GW: Yeah, WebAssembly was definitely a major component of the difference. There were a few others — how we approached sharding, how we approached the consensus algorithm. And the Polkadot protocol solidified and crystallized as we actually wrote it and implemented it.
The big difference was in what we call the shards statement the state machine of the shards and Ethereum's shard state machine was… I presume, based around the EVM. But please don't quote me on that, I haven't looked at the design for a long while — ours was very much based around WebAssembly,
So it's this idea that you can write any computer program that you want in any language that can compile to WebAssembly, which opens up a lot of different languages and you can really just write the blockchain, but with the idea that the blockchain will be secured by Polkadot. So it will fit within the same security model as Polkadot, it will fit within the same staking capital pool as Polkadot, [with] the same amount of stake is at risk if someone or a group coordinate and attack against your chain as if they would coordinate the attack against Polkadot… So a shard really is as secure as Polkadot, but it can do something entirely different and it operates independently — which is to say it operates in parallel, simultaneously work can happen on each of these shards...
So the key difference, I guess, in terms of the product, is that with Polkadot, you get the entire shard… Whereas with Ethereum, it was always about trying to take the Ethereum model and make it scalable, and the scalability generally comes from having these different ‘work tracks’. So [with] the different shards, having some smart contracts executing on one, and some smart contracts executing on another. And… the thing that really kicked off Polkerdot was saying ‘well, what if… these shards… don't have to be smart contract chains… what if there could be any kind of chain? What if one of them was specialized to handle a name registry, [and] one of them was specialized just for DeFi, [and] one of them is specialized just for NFTs’? Rather than making them all smart contracts, does this have an advantage?
And it seemed to me that there was a definite argument that it would have an advantage because there are some very high-throughput applications. Sometimes you just know that you're going to need to do a lot of transactions for this one particular application, for example, DeFi, and it makes sense to be able to have this specialization because you get a huge amount more performance through it. But furthermore, you get the ability to experiment, so you can actually try different ideas with different changes — you're not stuck [with] one single smart contract model, there are different ways of doing even smart contracts.
This idea that they're Turing-complete [is] kind of true, [but] there's a few caveats to it being true. It's not really exactly true, but even though it's Turing-complete, there are a lot of different ways of designing a smart contract system. And you can see this because there are several different chains out there now that have pushed forward different designs for smart contracts. Some of them work with[out] being Turing-complete; they work on languages that are guaranteed to halt after a little while. Some of them have storage fees, some of them don't have storage fees. There are all sorts of different things that you can do, and experimentation, is in my mind, what made blockchain great. It’s an experiment. It's a new thing and let's try this new thing, and maybe lots of different teams are gonna try lots of different new things.
So with Polkadot, the other thing that… we can try [is] if you have a hundred different parachains — a hundred different shards — we can try a hundred different things at the same time, and they're not just tied into this smart contract environment [for which] design decisions have already been fixed, but they can make their own design decisions. They're at the very top level, they're blockchains, they can make their own design decisions from the point of view of a blockchain. And this experimentation platform was really what made me excited.
CR: Got it. Is there a limit to the number of parachains on Polkadot?
GW: Almost certainly, yes, but we haven't found it yet. So we're going to keep rolling them out and as a limit comes along, we will optimize and take whatever action is necessary.
But one of the limits that we assumed early on was to do with message passing, so basically, as you add on more parachains, then there are more parachains that each one might want to talk to. So it's like what you add on the second parachain then they can only talk to one of the parachain — there’s only two in the system, ‘A’ can talk to ‘B’, and ‘B’ can talk to ‘A’. But when you add on the 101st parachain, then now every parachain… can talk to a hundred other parachains, which means if you think of it in terms of the total number of connections, then there's now 10,000 connections because yeah, because it's a multiplier… and this is commonly what's called ‘network effect’. It's the same reason that Facebook had this snowball effect, because, for every extra person, it became more useful… When there's a network of two people, then why would you ever bother joining, but [when it’s] a network of a billion people, then it's like ‘well if I want to join any network, it has to be this one because there's no other network that connects a billion people’.
So yeah, [regarding Polkadot], basically this message passing becomes… more expensive because it [is multiplied by] the order of the square, of the number of total chains. As you add an extra chain on it becomes much more costly to pass all these messages… We have an algorithm now so we have a way of doing a design that allows us to avoid this issue, and so actually we are much more confident that we can actually go much higher than our original expectation of one hundred and it may end up being in the hundreds. But realistically, there's going to be some limit, and we will work our way towards it and have other designs that will help us scale even beyond this limit.
CR: So the limit comes from when it starts to become too expensive for all the different parachains to message between each other, and you expect that to be within the hundreds when that limit comes. And a hundred is a good [estimate for] what Polkadot can support?
GW: That's what we have said in the past, and with our new design, we expect it to be more than that. So it may even be 200, 300, 400, 500 — we don't know quite yet. And realistically, it's actually possible that the bottleneck won't be with the message passing, but will actually be with the overall finality algorithm.
So we have this finality algorithm, it's called “Grandpa,” and for it to work, it needs all of the validators to basically talk to each other. So all the validators have to talk to all the other validators. and they don't have to talk very much. It's very simple, the talking that they're doing.
But nonetheless, all validators have to talk to all others and we have this, [and] the number of validators has to be 10 times the number of parachains. So this is where you get the problem. So suppose you have one million parachains, now you have to have 10 million validators, which means all 10 million validators have to talk to 9,999,999 other validators. So you suddenly have to have this huge amount more traffic for each validator. Every validator now has to have this vast amount more bandwidth, and that reduces the decentralization of the network because it basically means validators have to have very specialized equipment sitting at a data center with huge amounts of bandwidth. So we don't really want to end up effectively centralized in that way, like some other protocols which are willing to trade off decentralization. We don't want to trade off decentralization. We consider it very important.
And so for Polkadot, [we will] end our experimentation [and] end our validator increase at the point where we feel that the bandwidth required for any given validator is getting to the point where it would be unreasonable for any normal actor to be able to provide. But overall, if you want a number, then I would say somewhere [between] maybe 100 to 250 is really where we're targeting right now.
CR: Great. Just to take it to a higher level, the reason why these parachains or shards are important is because it gives developers the flexibility to build decentralized applications in the way that they choose to and because it allows Polkadot to scale, right?
GW: Yeah. that's right. So scaling is an important one and the scaling that we do through these parachains, these shards, is twofold. One of them is we've split the workload, so it's like instead of having just one person who's doing a bunch of work in an office, it's like they can hand it out to 100 different people and they can go away and do the work in their own offices and then come back and hand the work back to the original person. This is a very simple concept, parallelization.
The other way is by these shards being what we say “domain-specific.” So basically it's like the shards are built to perform very well at doing a very particular kind of work. So it's like, you wouldn't give your accounts job to a doctor to do because doctors do medicine and accountants do accounts. But the idea is that, with Polkadot, you can have a shard that is great at doing accounts, and another shard that's great at handling NFTs, and another shard that's great at doing reputation stuff, and another shard that's great at doing governance stuff right. And these are domain-specific, which means they can get through their workload faster. So we get scalability through the performance gains, which are delivered by not being general — so a platform like Ethereum is very general through its smart contracts, it's like a Jack-of-all-trades but a master of none.
So the idea was to say, ‘well in Polkadot, we can have some Jack-of-all-trades chains, sure, but let's have some master chains as well, that are really good at doing some particular thing very performant at doing some particular thing’. So the scalability was very important.
But also… when you make new technology, you only come to realize some of the repercussions later on, after you've built it and are using it. And one of the things that I came to realize later on with Polkadot, is that with smart contract platforms, they are perfectly good for doing smart contracts, but a lot of what is happening on these smart contract platforms is not really a smart contract workload.
A lot of the users of Ethereum fall into two categories. One of them is people who want to deploy… a decentralized application right? So they have this idea of a decentralized application, what Web3 application they may want it to be, they built it using smart contracts, and they launch it — that's one user-set. Another user-set is the people that want to use that decentralized application. For some, the issue is with smart contract platforms like Ethereum. The users don't want to use Ethereum, they don't care about Ether — they want to use the decentralized application.
A lot of the users of Ethereum don't want to use Ethereum, and if a decentralized application can come along on a platform where the user doesn't need to worry about the underlying platform's token or any of that stuff, then that's going to be better for the user because they don't care about the underlying platform of the decentralized application that they want to use. Similarly, decentralized application providers don't especially want to have to work their decentralized application into several smart contracts. Smart contracts are not necessarily… how you want to architect or implement your decentralized application.
Oftentimes smart contract environments force certain design decisions on you that you really didn't want to make, that are against how you would normally implement [and] architect software. And with Polkadot, we want to return that entire design spectrum back in the hands of the decentralized application developers. And furthermore, we don't want to intermediate Polkerddot, we don't want to get in there between the decentralized application and their users because that's silly. That's just reducing the utility of the decentralized application for the users because if the users had to care about the DOT token, or had to care about how to interact with Polkadot, or any of the other stuff to do with Polkadot, then it would be a distraction from them using the decentralized application.
Furthermore, it really just means as soon as you intermediate a platform between the application that sits on the platform and the users of the application, you're effectively restricting the users of the application to a large extent. To the users of the platform... it's really hard to use any smart contract on Ethereum if you don't already own Ether because, otherwise, you have to discover not only the smart contract, but you also have to discover Ether — you have to buy Ether, which means signing up at an exchange, funding the exchange doing any ID requirements of the exchange, getting Ether, sending Ether out from the exchange into an account, and then presumably doing something on the decentralized application from the account. That's a huge amount of effort to go to for someone that doesn't already know Ethereum. Realistically, your target audience is going to be people that hold Ethereum anyway, roughly know how to use it, maybe they’ve got Metamask in their browser, they understand the concept of an Ethereum account — and that's a big problem if we want to hit the mainstream.
So Polkadot is really about ensuring that it's delivering… a decentralized application platform to developers and allowing them and their application to interact with the users directly without having any intermediary involved.
Polkadot’s parachain architecture
CR: Got it. So for example, users of Acala — an application built on Polkadot — don't require DOT to pay for transaction costs; they can go and buy the ACLA token directly. So where does the value of DOT come from?
GW: So we don't want to economically disconnect DOT from the success of its parachains, it's quite the opposite. What we want to do is ensure that the users of the parachains, of these decentralized applications, don't need to care for their day-to-day activity. So what we do is we say ‘a parachain can deploy on Polkadot if it either is a common good parachain’, so basically a piece of functionality that is useful in and of itself to DOT tokenholders — so that's one way. But assuming that's not what's happening, then we have these auctions.
So the idea is that we auction off these 100, or 200, or whatever, parachain slots — so these little blocks of sort of ‘Polkadot land’ to teams. And teams can bid on these slots, and if they win one, then they get to deploy their decentralized application into Polkadot — and that's it, and then they can appeal to their users in whatever way they want without the worry of having their users having to be exposed to the DOT token.
So these auctions happen as a lockup. So there are no DOT tokens paid out by the teams, they don't have to raise a large portion of DOT tokens and then hand them to the protocol. Really, it's about locking them up, and their DOT tokens are locked up for as long as the para chain team holds this lease this slot. At the moment, that's a two-year lease for Polkadot. So basically, the parachain team agrees to bid… DOT tokens and then these DOT tokens are locked up for two years, and two years later the team gives up its parachain slot and it gets [the DOT] tokens back again.
Now, the interesting component to this, is how is a team going to go to find [large sums of DOT] tokens to put a bid in for a slot? Maybe they'd have a lot of DOT tokens — that's a possibility, but if they don't, then what do they do? Maybe get a loan from maybe a VC or someone that wants to support them… It's important to say that the tokens are guaranteed by the protocol… they will definitely get their [DOT] tokens back again — that's not a risk.
The other way of doing it that we introduced, is this concept of crowdloans. So it's a bit like crowdfunding, except each of the people in the crowd only loans their DOT tokens — they don't buy anything, they don’t give anything, they loan their DOT tokens. And again, this is trust-free — it's part of the protocol, which means they are absolutely guaranteed to get their DOT tokens back again. And this allows a team to basically say ‘hey, we think our parachain is going to be great, you should support us’, and allows them to go out and try and get support, and have people put their DOT tokens in, and maybe help them get a big enough bid to get a slot in the auction. Now, the team might say, ‘well, if you help us get a slot, we will give you some sort of reward’. And so, you end up with these parachain teams offering deals whereby people who loan them enough DOTs for their parachain slot get some tokens back, or whatever else.
But interestingly, the reward can be fairly free-form, so at the moment we’ve seen rewards that have basically been ‘we'll take 5 % of our token base and hand it out to people that have helped us get a slot’. That's fair enough, that's one way of doing it. There are others — one of them might be where you just give an NFT which allows you to become a collator — being a collator is basically like a validator — gets you some tokens for every block that you validate. So [rather than] handing tokens out immediately, you can hand tokens out over an ongoing period. Or you can use it as a key to unlock some functionality within the chain — for example, becoming a collator or some other keeper role that the chain might have. Or you could maybe make it so you get a particular reduction in the chain services — there's all sorts of ideas that are out there that chains might choose to implement.
CR: Got it. The value [of DOT is] driven from teams who want to participate in the ecosystem and build on parachains, not through the application users themselves. So far, how many parachains have been auctioned? How much are they going for?
GW: On Polkadot, we have six active parachains. The biggest one? I'm not actually sure, I think… it was like 1% or in the single-digit percentages — maybe 1.5% to 2% of total DOT issuance per slot for the initial ones. Although I’m really terrible with the financials.
CR: Okay, so these teams are raising about a million or so DOT tokens and staking them in the Polkadot ecosystem, and then when they want to leave or when they're done with renting their parachain, then they can take those DOT tokens back?
GW: Essentially yes.
CR: Okay. And then you're saying that the teams can either buy these DOTs or borrow them from investors or from the community itself, who will then get something in return — a yield in tokens, or maybe a discount in whatever application they're building, or something in return further for the crowd loan.
Polkadot and Kusama
CR: Okay. Then there's also Kusama, so I'd like to also get that breakdown. There are so many different terms that I want to clear up.
So there is substrate Substrate, which I understand is like a framework for building parachains. Then Kusama is like a separate relay chain? But yeah, can you explain how Kusama differs from Polkadot itself? And I also understand there's like a bunch of parachains on Kusama, so how do all these different pieces within the Polkadot ecosystem integrate?
GW: That’s right. Substrate is a parachain framework, so it's like a toolkit for making chains. Interestingly, it's not just for parachains, it's also used to make Polkadot and Kusama themselves. So it's able to create blockchains of any kind, and it's designed to be super modular so that we can do all of the various interesting stuff that we want to do with Polkadot, and we can evolve it and bring in different consensus algorithms, and different ways of handling shards, and all this kind of stuff down the line.
With regards to the difference between Kusama and Polkadot, they're both relay chains, so they're both fairly similar to each other in terms of technology. Kusama is what we call our “canary network” — which basically means it's where we try out new technology and new ideas before we think we’re ready to put it on Polkadot.
Polkadot is a high-value network and we don't want to risk the possibility of deploying new technology into Polkadot and having it break or something. Although we do a lot of testing before it ever gets anywhere close to Polkadot, at the end of the day, these are very complex systems and it's very difficult to predict with certainty how some change is going to affect [the network], and so we deploy it to a system of comparable complexity — Kusama, in this case — with some substantial economy behind it in order to try it out and see how it works.
Now, there are two classes of new things that we try on Kusama before we try it on Polkadot. One of them is just general new technology. For example, the parachains’ code — this was all tried on Kusama before it moved over to Polkadot. And Kusama actually has way more validators than Polkadot, I think more than a thousand validators now. Generally, the people who are the validators for Kusama are running on not such good hardware compared to Polkadot validators, just because less is paid out for the Kusama validators. So we can really get a good feel — if it works on Kusama, then it's probably going to be okay on Polkadot.
The second thing that we try out on Kusama are ideas that may never even make it to Polkadot, things like ideas for governance. So governance is a relatively new frontier for blockchain and there are a lot of interesting ideas that I and others have for trying to get the wisdom of the crowd, or the wisdom of the market, or whatever, involved in the decision-making for blockchain — rather than default to this benevolent dictator that Ethereum seems to have got into, or this utter shambles that Bitcoin has sort of gone into. We really want to get good governance.
There are a lot of different ways that we can bring people together to get decisions made, and all sorts of different voting mechanisms, all sorts of different government bodies that you might want to introduce for checks and balances, for making sure some things go faster, and making sure certain things take more time. So there are a lot of different areas of experimentation, a lot of different experimentation, a lot of very big design space in governance systems — especially now that we have Blockchain that helps us do things that would otherwise not be possible when we have to use more in-person systems, as oftentimes in the past, governance was constrained. So Kusama provides a really good platform for trying out these kinds of new ideas, like governance structures or social structures for decision making. One of them was the Kusama Society. This is a piece of functionality that's on Kusama, it never made it to Polkadot [and] probably will never make it to Polkadot, with the idea being — what social structures can we build that can do things that we wouldn't otherwise be able to do?
So Kusama facilitates experiments in two key directions, one of them being ‘let's make sure that this isn't going to break Polkadot’, and the other one being ‘let's really push the boat out and see what we can do in this particular direction of technology’.
CR: Very cool. So do you envision, within the Polkadot ecosystem, basically there only being these two chains — Polkadot… as a mainnet with all its different parachains, and Kusama as like a test net — or will there be multiple different relay chains that may use substrate?
GW: I suspect that we may see one or two others. The ones that we are seeing at the moment are consortium chains, so chains where a particular industry has a bunch of individual organizations, companies, whatever — they want to come together and interact in some way, but they don't necessarily want to be part of a public network, maybe they don't see the reason why they should join another network's economy. For this, there are good arguments for why Polkadot is a more reasonable proposal, because it doesn't try to intermediate itself in all the transactions. But in principle, it may be that we find that relay chains that ultimately are bridged into the Polkadot relay chain may become part of the wider Polkadot ecosystem. I think it's fair to say, at least speaking for myself… that we’re not really going to see… relay chains that don't provide any key advantage over Polkadot or Kusama. They would just be some sort of standard copycat thing and they will probably exist, but they're not going to be welcomed with open arms.
One blockchain to rule them all?
CR: Within the place where we are in the blockchain industry, where multiple Layer 1s are competing against each other, where do you see Polkadot? Do you believe Polkadot will become the main blockchain ecosystem that will attract most developers and users? Or do you see a place where all of these different Layer 1s will interconnect with bridges? Or maybe another option? What's your take?
GW: There was an article put out a few weeks back by Vitalik basically saying ‘bridges are broken, don't do stuff on bridges, and broadly speaking I'm in agreement. I think bridges probably have a fairly small role to play, but overall I think that they suffer from what's called the inverse network effect — which basically means they're secure enough for light amounts of financial traffic, but they are definitely not secure enough if they turn out to be very important and traffic grows, then they become increasingly insecure because the security of the bridge is basically fixed, but yet the value of the traffic can grow very substantially, and at some point, the value of the traffic outweighs the security of the bridge and at that point, it's only a matter of time before someone attacks it. Bridges may well be reasonable as long as the value of the traffic is kept reasonably low, so basically as long as the bridge doesn't get used too much. But to architect an ecosystem [based on] bridges would be asking for trouble. So yeah, I don't see that as being a viable way forward.
I would hope that we retain, to a degree, at least two or three different major chains and different major architectures. We always talk about Polkadot as a Layer 0, that it doesn't really do anything itself beyond secure other chains, secure its shards, secure its parachains — these are the things that provide applications to users. So I think it's not far-fetched to imagine a world where there really are just one or two Layer 1s that can just about secure themselves, and then a few or maybe even just one Layer 0 that is securing the rest of the blockchain ecosystem. But I don't know.
My main driving factor at this point is to ensure that we provide the very best platform for developers, because in my mind it's not about user acquisition at this point, it's about developer acquisition and ensuring that we have the most compelling applications and that we have the most compelling platform for applications to be deployed onto. It's very important in my mind that decentralized application providers can effectively market their service to individuals without having to get them to buy into the platform that they happen to put their service on — this in my mind is crucial for mainstream adoption, and really what I'm focused on with Polkadot is making sure that that is the reality for Polkadot because it's not the reality for any other chain.
And that the APIs and the overall framework is good to work with. And we'll probably play around and make a few little test applications, demo applications of our own because firstly it's fun and secondly it's a good way of testing out your own stuff and making sure that it's good.
CR: Okay, going to the original question, in this world where there is one Layer 0, which is Polkadot, and then a few Layer 1s — would these Layer 1s be parachains?
CR: That’s interesting, so in that future Polkadot just wins, and there are not many other blockchains that are being used?
GW: I would look at it like Polkadot is a security alliance at the end of the day. It's a way of allowing the different parachains to come together and give value through their own utility to an overarching token, the DOT token, in order that they be secured because they gain their security from Polkadot's consensus. That's the value of DOT, really just providing security and messaging between the chains.
So if we look at it in that way, does there need to be more than one security system? To take a very timely example, does there need to be more than one NATO? I mean, if the purpose of it is basically ‘join this and you won't be attacked’, and it seems to me that there's no reason not to join it. If it's fair and transparent and open, assuming that it fulfills what it says on the tin, then it seems to me that there's really no reason not to subscribe to the security that… everyone else is under. If that's the case, the way of saying it is that security has network effects right — it's more valuable to be under the security system that 90% of the value is under than 10% of the value is under. And so similarly with messaging, there's exactly the same network effect. It's more valuable to be in the messaging service that 90% of the blockchain world is under rather than 10% of the blockchain world is under.
Messaging and security go hand-in-hand in the consensus environment simply because if you receive a message and do something and then it turns out that the sender of the message was not finalized, was not the final thing, or was in some sense corrupted, then you have a problem because you have already acted on the message and it's hard for you to wind yourself back. And it may be that part of your action was to tell someone else to do something, and now they're doing something and they can't wind themselves back, and it all gets very messy.
And this kind of messiness where there's an initial problem and that problem sort of infects something else, which infects something else, which goes on, and on… and infects the entire wider consensus system — this is a similar setup to the thing that caused the 2008 financial crisis. It's this idea that you've got this initial bad asset that's coming out of wherever, the U.S. sub-prime property market, and it's basically wrapped, given to someone else, someone else then wraps it up and gives it to someone else, then before you know it, everyone is building huge amounts of financial derivatives on top of this thing — everyone's holding it, it’s a huge problem, and then you find out the initial thing was bad and now the rug is pulled from under the entire wider industry. This is basically the case if you try and connect blockchains without having shared security. The rug can be pulled when one that sends a message turns out to be a bad one.
CR: That makes sense. So do you believe that Polkadot can become secure enough and decentralized enough to support global economic activity? Do you see a future where all the financial system, all of the web apps, Web3 — everything is running on top of Polkadot? Or do you see it as a more limited future where there will be some blockchain-based applications but not the entire internet and not the entire financial system will be running on blockchains, and that main blockchain that's being used is Polkadot?
GW: I think we are in a period of a semi-golden era of blockchain at the moment in that there are a few different directions that are being explored, and I don't think that level of experimentation is going to continue indefinitely.
I find it hard to see there being more than a few very high-value chains, so chains that secure themselves. I think that's because securing yourself is costly — it requires a lot of capital — and kind of pointless if you can do it some other way — it's just cheaper and easier to do it some other way.
So I don't think that we're going to see very many of them, and the ones that we see will likely each do different things. They'll have different design trade-offs that are each valuable in their own right. It may be that a very scalable, actually decentralized, smart contract chain has its own security for the foreseeable future, maybe. But I think, thus far, Polkadot is the only chain that really provides this parachain product which is basically the ability to secure any other chain — that is novel and I think it's actually very valuable.
So if we think about global levels of transacting, like I'm not really sure at what point we decide to go to Layer 2 and say ‘well, we don't really need all of the security that is under Polkadot, or Bitcoin, or whatever’. For example, do we really need that level of security to count the number of likes that a particular post has? Maybe maybe not — arguable, definitely arguable. But at the moment, we're aiming for Polkadot to have the order of about one million transactions per second that it can support under its primary consensus, basically with the same level of security as the Polkadot relay chain.
We already have designs for systems that would be kind of “Layer One-Half,” in Polkadot’s terminology, basically chains that are roughly part of Polkadot's overall consensus, but they have a separate consensus of their own which gets eventually folded into Polkadot's consensus. So there's kind of like a sub-consensus, or like a mini-consensus, with the idea being that we could get a huge amount more transactions through on these. So we go from maybe one million transactions per second to maybe [from] 100 million up to one billion, and then from there, potentially even further.
Aiming over the next 3 to 5 years, having in the hundreds of millions of transactions per second is where we should be looking while retaining decentralization.
CR: So does that mean that you do see Polkadot being able to sustain global economic activity?
GW: I haven't actually taken the time to measure how much that would be, but it's certainly way higher orders of magnitude than Visa would be able to process. So if Visa is any kind of yardstick then yeah, we're definitely looking to [exceed] that by maybe one-thousand times.
CR: Got it. And then to wrap up, I wanted to wrap up Web 3 because this is something that you've been talking about for a very long time, probably before most people had heard of the term — you wrote a piece on Web 3 in 2014. If you can just go over what Web3 is for you?
But specifically, what I really want is the criticisms of Web3 and how you look at them? Recently, there's been this blowback or pushback from large groups of people who are against NFTs and think that web 3 is unnecessarily financializing activity online. Also, tokenized governance has turned web applications into plutocracies. So unnecessarily adding [things like] greed and money to the internet is the common criticism of Web3. So I wanted to hear your thoughts on it given that you've been thinking about this for a while.
GW: So Web3, for me, is not cryptocurrency, it's not blockchain, it's not tokenomics. Web 3, for me, is decentralization, it's openness, it's transparency. And it's the ability to have absolute technical guarantees on what your expectations are, and ensure that they be met. So I should not need to trust Facebook in order to send a message to my friend. I should not need to trust my bank in order to send money to my father. That should not be the case. I never used to need to do that.
It used to be, assuming that my father was next to me, that I could transfer money to him by giving him a £10 note — there you go, there’s some money, that transfer happened. I didn't need to trust any bank to do it. I can talk to my father, my friend — if they're right next to me, I don't need to trust Mark Zuckerberg. This idea of building trust into everything is lazy, it comes about from laziness. It's because it's easier to architect something if you give yourself the right of god as the developer, as the service provider. It's harder to do it when you don't, but that's nonetheless what we should be doing in this era, where we have very powerful people in the world that are doing their best to try to kill off the concept of a free society.
We have no choice but to enshrine these guarantees of privacy and these basic rights into software — into the very technology that we use — so they cannot be worked around, that no individual can be leaned upon to break these rules. This is crucial if the free world is going to survive. That's my thought. And Web 3 is a technology set that allows us to do these kinds of things — these massively multi-user applications, these web applications, these internet applications. It's a set of technologies that allows us to build internet applications that are not susceptible to attack from those that would rather not have a free world.
CR:I just love that answer. Okay, last question, how are you defiant?
GW: How am I defiant? I um…
CR: Or maybe you think you're not defiant? But I think you are.
GW: Well, I guess it takes a lot to stop me from continuing on the path that I want to continue. I have a pretty clear idea about what it is that I want to build, and if people come along and say ‘NFTs are bad therefore you shouldn't build Web3’, I have no problem defying them.
I suppose I'm defiant in so much as I have my beliefs about what is good engineering, what is good technology, the way that a good society should be in some sense, right and wrong. Like I don't think the U.S.'s war in Iraq was particularly right, and I think that a lot of crimes were committed that never got punished and I think that is wrong. And I don't want to single out the U.S., most countries have their skeletons in the closet — some more than others. And I think that openness and transparency can be a great crucial tool in ensuring that society stays free and that, in some sense, the bigger rights and wrongs of the world become known. And I'm happy to defend my opinions on that, and I suppose in that sense I can be quite defiant.
But I don't know, I prefer to build and I'm not really out here to cause controversy.
CR: Love it. I guess your high convictions about openness, transparency, and decentralization can make you defiant. With your talent, at least you will build to see those values play out.
Awesome! Gavin, it was a pleasure chatting with you. Thank you so much for extending the interview a bit longer, it was really valuable hearing all about a Polkadot from the man himself. Thanks again for joining me! Again, really a pleasure!
GW: Likewise. Lovely to chat, Camila, anytime.