Joe Petrowski: 00:00 Welcome to Relay Chain, a podcast produced by Parity Technologies where we discuss all things Substrate, Polkadot, and Web3. Today, we have Gavin Wood for the first episode of Relay Chain. He's the co-founder of Ethereum, the founder of Parity, and founder of Polkadot. Welcome. Gavin Wood: 00:17 Thanks. Joe Petrowski: 00:17 I want to start with the early Web3 philosophy, and I read that prior to Ethereum you were working on music and legal software, but on a 2014 podcast you said that Ethereum would just be a weekend project, or over Christmas break, but you also had some blog posts that were outlining Web3, like from low-level messaging, to frontend stuff. How much had you thought about this? Was the Web3 vision something that came really quick, or was it a really long time in the making? Gavin Wood: 00:47 It was a while, I guess, before ... I think it was I first sort of rediscovered Bitcoin in like 2013, and it wasn't long after sort of understanding it better, how it worked, that I sort of came to the conclusion that for it to survive it was going to need to decentralize not just the basic currency itself, but also things like the exchanges, so the infrastructure that was, at the time, entirely centralized. Gavin Wood: 01:15 And I came up, initially, I guess it was like May, June time, 2013, with a design about how it would work. A sort of decentralized message passing, with order matching system. It didn't really go anywhere. It was just sort of a thought experiment, just to sort of imagine how a decentralized application, like an exchange application, might work. Gavin Wood: 01:42 But fast forward to like March, April, 2014, and it became clear that Ethereum was going to suffer from a similar sort of problem in that you can decentralize the very core of the protocol you want, but if you don't have decentralized infrastructure surrounding it, then it's not going to fulfill the overarching purpose. Gavin Wood: 02:06 And I guess Whisper, as the communications sort of mechanism for that Web3 platform that I envisioned back then, was pretty similar to the order distribution architecture for the decentralized exchange thing that I'd come up with a year before. Gavin Wood: 02:27 The other main main component of Web3, as I envisioned it back them, was the Swarm component; this idea of a decentralized publication mechanism. And really, that was based almost entirely upon BitTorrent, plus Kademlia, as a means of basically getting, once you have a hash, you'd be able to find the pre-image to that hash. Gavin Wood: 02:51 Someone would have it somewhere, and it was really just a protocol to locate it and download it. The idea behind sort of Swarm, Whisper obviously grew beyond these two basic origin points. With Swarm, eventually, I imagined it to be much more sort of cryptographic, with the idea that no one would, by looking at your traffic, know precisely what you were trying to download, and the same with Whisper. Gavin Wood: 03:22 No one would know, looking at your traffic, who you were communicating with, and when you were communicating, and for these to really be delivered I think we're going to have to start looking at more additional technologies like mixed nets and so on that might be able to help. Joe Petrowski: 03:39 Yes. I don't hear too much about Swarm and Whisper right now. Do you think the underlying goal stays the same? Have they been replaced by other technologies? Gavin Wood: 03:49 Well, Swarm and Whisper were never really specific technologies. They were rather kind of placeholders in this platform. Web3's sort of a meta platform. It's like a set of technologies which, taken together, can help people create decentralized applications, and distribute them, and so on. Gavin Wood: 04:08 But it was never really meant to be a singular platform that moves along with particular versions. In principle, Bitcoin and BitTorrent could form part of a sort of notional Web3. Its not to say that these are not components, the sort of pieces of the puzzle. It's just that, as time goes on, I expect that the technology will develop, and we'll get increasingly closer to being able to deliver this idea of a decentralized application platform that's massively multi-user. Joe Petrowski: 04:42 Yeah, so you started out by talking about a lot of the interactions between a technology like Bitcoin, and the outside society, as kind of the exchanges. How do you see this interaction growing from just an exchange, into a business, into maybe like a nation state, or more social governance? Gavin Wood: 04:59 I guess it's kind of interesting in that I don't see these decentralized applications really being obviously more comparable to a business, or a nation state, or a protocol than any of the other possibilities. I won't say Bitcoin is more a nation state than a protocol, but I wouldn't even really say it's more of a protocol than a nation state. I wouldn't say it's anymore of either of those two than a business. Gavin Wood: 05:28 It's none of the three and it's all of the three. It's a new kind of type of thing that we don't really have a word for. I mean, we've got some terms like DAO and DAC and all the rest of it, but we don't really have a full understanding of what it is, because we haven't really seen it fulfill its potential quite yet. Gavin Wood: 05:49 But in general it's definitely going to have components of all of these three things, and so is any kind of decentralized application that's massively multi-user, so basically everything that we can imagine being on Web3 will fulfill some of the characteristics that these three things have, but I don't really see them as being closer to one or more of the ... against the others. I mean, you know, obviously some things, particularly use cases, will lend themselves more to replace, for example, a business than they will to replace a nation state. Gavin Wood: 06:26 But if we're talking sort of general about the technology, then no. I think it's really a new thing and it's kind of different to all three. Joe Petrowski: 06:33 Yeah, I don't mean so much a comparison to where it falls in line, but what are the tools to integrate it with these other constructs that we already have in the real world? And maybe a better way of framing it is just saying, like, you've written that Web3 technologies give you some verifiable guarantees like that the person who sends you the message is the message. Joe Petrowski: 06:54 I guess, from a ground up approach, what are these guarantees, and then how come these small set of guarantees can actually provide the foundation for something much larger? And why are these the set that are the most important? Gavin Wood: 07:07 Well, it's three questions. I'll try my best to do the first one. The guarantees are pretty ... I mean, you know, you've got the document probably there. You can probably spell it out better than I can, but the guarantees are really about the sorts of things that we take for granted, that they're kind of explicit expectations that users typically have that we, as people, typically have when we interact with these massively multi-user systems. Gavin Wood: 07:34 Basically websites, you know? Web applications. But they often go unfulfilled, and the reason they often go unfulfilled is because the systems by which we are interacting are kind of just little more than facades to companies, or people that we don't really have a ... we would like to think that we trust them, and it may well be that we do trust them, but we don't really have any real reason to trust them. Gavin Wood: 08:03 Now, you know, national laws and international law as far as it goes do try to help create some sort of trust, but in reality, the rationale for that, there's nothing there. It doesn't really exist. The only reason that we trust that Facebook isn't going to lose our data from another security fault is because, well, they haven't maybe lost our data so far, although of course they have, so we can't really trust them anymore. Gavin Wood: 08:27 But we tend to give these things the benefit of the doubt, and that's unfortunate, because a lot of the time they really don't deserve it, and there is nothing more than this kind of blind benefit of the doubt-level trust that they have going for them. The technology does almost nothing to protect our interests, and so really what Web3 is, it's saying, "Right, well, the technology, let's actually build this from the ground up so that our interests are fundamentally protected." Gavin Wood: 08:55 We get given these guarantees just by virtue of using certain pieces of technology, so yeah, knowing that who you're communicating with is in the open, we should assume that. We should assume that anybody who has access to an internet connection can know who that we are talking to, unless we take action against it, and if we take action against it then we should have the guarantee that they will not know who they are talking to with some substantial degree of certainty Gavin Wood: 09:22 Similarly to what we're saying, it should be possible to know that if we're talking to our friend over the internet then random eavesdroppers are unable to determine what it is that we are saying to our friend, because we expect these kind of one to one conversations, and it may be talking, it may be sending pictures, it may be doing video; whatever. We expect it to have some degree of privacy. Probably quite a high degree of privacy, at times. Gavin Wood: 09:48 All those expectations should be met in guarantees from the technology, not from companies like Facebook or like Google that are able to tell us, "Oh, yes, don't worry. We've got your interests covered. We'll take all of the precautions and make sure that it's not hacked by various nefarious actors." No. It should be protected by the technology itself, and as such it should be open, and transparent, and auditable so that we don't have to trust any individual actor, be it a technology provider or a company or a government that our interests are being looked after. Gavin Wood: 10:20 Rather, we can trust people that we already know and understand and are comfortable with the actually go ahead and audit them. We can place in the hands of impartial actors, or people that we have already come to rely on, to make sure that they also believe that this technology is correct, so to take a very simple example, it may be that my mother doesn't understand that this technology can protect her interests particularly well, but she can ask me and I have a much higher degree of understanding, and yet she also trusts me because I tend to tell her the truth about things. Gavin Wood: 10:52 Now, that's not the case if the technology is closed, and opaque, behind a company's profit wall. In that case, all we can do is trust the company, and then similarly, in certain instances, it's difficult to trust governments, and there are various sort of criminal actors and elements online that make it very difficult to trust anyone that you would generally sort of meet cold call style online. Joe Petrowski: 11:14 Yes. Certainly, like, the outcome sounds great, but you know, a lot of people who have power, whether it's states or corporations, are not going to see that gracefully, and so we've already seen a lot of banks, and/or Facebook, kind of offering these watered-down blockchains. Which aspects of the system do you view as being critical, that you can't really make any compromise on in order to get these systems in place? Gavin Wood: 11:40 Well, the vision is the vision, and I would set that as the desired end goal. Although these things generally have, in reality, there is a sort of degree of compromise that is simply because these sorts of ideals and principles are rarely actually pragmatic and practical, but nonetheless, it's something that I think we should all have in mind as being the ideal end goal of all of this, and I think seeing how the world's sort of realigning itself in the last 20 years, I'm not particularly worried that the world will eventually sort of realign and accept this kind of trust-free technology over time. Gavin Wood: 12:28 And yeah, sure enough, the large actors in the world at the moment, the establishment that sort of get by because we trust them or because we cannot help but trust them, we cannot help but use them regardless of whether we trust them or not. In fact, these guys are not super happy about the idea of being disintermediated and eventually having their relevance reduced substantially, because actually we don't need these big behemoths to tell us who and who we can't communicate with, what we can and cannot say, and et cetera. Gavin Wood: 12:56 Rather, we will have technology that allows us to be sure that our personal interests are being looked after, so I'm not particularly bothered about working out how much of this technology banks can use in order to increase their profit margins. I actually don't think it matters. I think what matters is ensuring that the technology is developed and delivered in order to make the current requirement to trust these kinds of actors, the banks, and the insurance corporations, and the governments, and the Googles, and the Facebooks, to make the requirement to trust them unneeded. Joe Petrowski: 13:38 In order to make it pragmatic for people to adopt this technology, there has to be a more user-friendly experience, and I see one of the hurdles in the user experience of blockchain or crypto technology right now is that you have a lot of responsibility. Like, you have to keep your private key safe, and do you see that web2 has an inherent advantage in UX? Or how do you see UX evolving, and what are the big hurdles to developing a better experience? Gavin Wood: 14:08 It's early days. What we're looking at is something that, you know, if we're software people, would say is pre-alpha. This is little more than a tech preview, so yeah. At the moment, it's like, "Oh, you have to keep track of your key, and you have to do all sorts of things you're exposed to language like signatures and transactions." Yeah, it's not user-friendly. It's not usable, accessible, but that's because we're still very much in the early stages of being developed, we're like the internet in '92 or whatever. Gavin Wood: 14:45 This stuff isn't designed to be used by normal people for normal things. It's important to build the fundamentals before you start relying upon them for important stuff. But nonetheless, you can start to see some initial sprouts flowering of usability and accessibility, even in sort of Web3 terms, and over time we'll transition more from traditional centralized models to sort of new, decentralized models. Gavin Wood: 15:18 And over time, as they become more popular, they'll become more accessible and usable, and people will develop an increasing amount of effort to ensuring that, for example, if you lose your secret key in one place that there will be simple and effective ways of regaining access to your various things, and there are ways and means of doing it. It's not that the ideas don't exist, it's simply that solutions haven't been developed, they haven't been engineered, they haven't been tested. They haven't been rolled out, delivered, and adopted, but over time that will happen. Joe Petrowski: 15:50 Yeah, I agree with you, we're at like '92 phase in the internet in this. But you don't see anything that's inherent in the Web3 backend that prevents you from having a good experience, or a comparable experience on the front end, or possibly even better? Gavin Wood: 16:06 I would expect it to be even better. Of course, there are certain types of things inevitably that will be favored, that will favor either a more centralized or decentralized architecture, but I think, in general, it will be at worse no different, and at best an awful lot better than what we have at the moment. Joe Petrowski: 16:27 Yeah. I want to transition a little bit towards governance. You and a few other people have talked about the internet as kind of a new jurisdiction. How does this mesh with the geographic concept of legality, and what gets preserved and replaced? Gavin Wood: 16:44 Legality is a concept largely deriving from the idea of nation states and sovereignty, so a nation state is sovereign or has sovereign laws that it can enforce and punish as it feels best, and in some very digital sense, blockchain is similarly sovereign. It has a bunch of laws or rules that it can enforce and punish as it sees fit, so we have two different kind of sovereignties here. Gavin Wood: 17:17 I mean, they're both very much all about enforcing rules, but the notion of the digital world being ... the sort of well-connected, fully globalized, completely integrated digital world being enforced with national level, which is a purely sort of geographical boundary, national level rules doesn't make any sense. Gavin Wood: 17:40 And we're already seeing how little sense it makes by the fact that, you know, we have VPN providers that trivially allow you to, if you don't like this particular website knowing that you're from the UK, then fire up a VPN and they'll instead know that you're from Latvia. Suddenly, national barriers don't mean anything on the internet. Gavin Wood: 18:01 And you go to China, and they mean a little bit. It's a bit more of a pain to run a VPN in China, but it's still possible. You arrive at the airport, you can fire up your phone, fire up a decent enough VPN, and sure enough the great firewall of China is not so great. I don't see that changing. If anything, I see it sort of continuing in the same vein. I don't think that national barriers are going to become a huge problem. Gavin Wood: 18:26 Now, there's a small chance that the UN or something forms some internet standardization firewall thing that means, no matter where in the world you go, you still won't be able to sort of access things, or whatever, use the internet in a certain way. I think the chances of that are very limited, very small, primarily it's technologically very difficult. Even in China, it's still possible, even without VPN, to sort of find your way through ports, and the dark web or whatever, outside of the material that China would maybe prefer you look at. Gavin Wood: 19:04 I think, technologically, it's not possible. I think also just simply by the fact that national governments tend to be in one level or another of competition, that means that they will find it almost impossible to agree on any kind of global firewall, and as such I think that we will end up with a world where the internet is its own jurisdiction, or rather set of jurisdictions, and blockchains will become sort of their own sovereign nations within this world. Joe Petrowski: 19:40 I guess one of the implications this will have, if you're involved in the governance of a blockchain, but you live in a particular country, how do these things collide or hopefully not collide? Gavin Wood: 19:53 An interesting question. We don't really know, because blockchain governance is so novel. There isn't really much to go off at this point. If you are physically located in a country, then whatever it is that you do, they will consider their business. If they know for a fact, subject to their own sort of laws, if they can demonstrate in some sort of court of law that you are behind such and such an internet identity, then yes, they may well hold you to account for whatever it is you do as part of governance, and depending on that particular country, you know, freedom of speech and whatever rights that they give you, it may be an issue that you are taking part in governance. Gavin Wood: 20:45 But I'd say that's a fairly big if, and I'd say it largely ... it's more for maybe actors that have very critical roles in governance, but for decentralized governance where we have a multitude of much smaller actors, and we see the emergent effect in them coming together, I find it hard to imagine any particular nation state going after these relatively small fish. Gavin Wood: 21:15 I see it, in fact, as kind of similar to let's say the governance of Bitcoin, yeah? Not that there really is any governance, but if we want to get relatively low level governance, then we can imagine it's Bitcoin nodes, and miners, and transactors signaling their intents, and is there ever a sort of nation state that have said, "Right, you're running a Bitcoin node or you're mining, therefore you're contributing to the DAO that is Bitcoin, and therefore you are helping run this alternative currency." Gavin Wood: 21:50 Which maybe the country has decided is wrong, is illegal, is breaking some sort of money transmission laws. No, we don't sort of see that. It's assumed that if enough people are sort of doing it then it's basically going to be fine and the government will just have to get on and do something else. Joe Petrowski: 22:07 So when it comes to governance on-chain, a lot of protocols integrate some form of coin voting. Polkadot introduced a council kind of centered around minority rights, and then the fact that money controls modern democracy not withstanding, one of the goals of liberal democracy is protecting minority rights. Can you expand on the difference between coin holder voting as a narrow topic, and on-chain governance as a bigger picture? Gavin Wood: 22:37 Sure. Coin holder voting is basically this idea that if a decision needs to be made within some system that has coins, then each of the owners or controllers of those coins will have some voice in this decision, and the amount of voice that they have will be generally proportional to the amount of coins that they have, so one coin equals sort of one vote, so to speak. Gavin Wood: 22:59 And normally there's a majority carries assumption. There may be some additional quorum requirements as well, but it's a very simple mechanism. Now, on-chain governance in general is, "Forget that particular mechanism." And it's the idea that on some consensus of blockchain, usually, there is a decision to be made, and on-chain governance, in my mind, at least, also implies that when the decision is made, it's enacted. Right? It's enforced. Gavin Wood: 23:35 But regardless of that, we assume that there's some decision to be made, and there is some mechanism on the chain that the chain itself enforces, some process that the chain itself enforces, which determines what the outcome of that decision should be, what should be decided. Gavin Wood: 23:54 Now, how it determines it, we leave undefined. Right? That's the magic of on-chain governance. Different chains may have different types of governance. Some chains may do this stupid sort of coin voting thing. Fine, whatever. It's pretty dumb, but you know, it's the first thing, so whatever. You know, you got to start somewhere. But other chains, you know, we would hope, like Polkadot, introduce other, more sophisticated, more nuanced mechanisms. Gavin Wood: 24:23 And eventually I'd hope that the on-chain governance sort of proceeds, and we end up with some kind of working futarchy kind of model where the mass, the crowd as a whole, feeds in information under some sets of economic interests, that mean that the chain will ultimately be able to make whatever the correct decision is based upon some fundamental assumptions. Gavin Wood: 24:50 You know, whereas Polkadot is certainly not the futarchy level, there is some sophistication behind it. We do introduce concepts like lock voting, which means that it's not a simple coin holder vote. Rather, there's some degree of conviction. If you are happy to lock yourself into the system, and the potential ramifications of your decisions for longer, then it means that your opinion counts more. And I don't think that this is certainly the be all and end all, but I think it is an important step towards this ultimate desire that blockchains should be able to govern themselves so to speak, and I don't mean that, again, blockchain, sort of the computer rules, but rather it shouldn't be the case that individuals rule, as is the case in sort of traditional governments where you elect a president or a prime minister, and you have a cabinet, and they basically make decisions over certain things. Gavin Wood: 25:48 It's people who make decisions, but rather that a blockchain makes the decisions based upon combining all sorts of inputs and factors, and some of those will be the opinions of coin holders, and there may be others as well, in such a way that no single actor or no small set of actors can influence the vote in some nefarious way. Joe Petrowski: 26:10 But also the ability to change the factors and inputs and rules by which it updates itself. Gavin Wood: 26:17 Sure. That will be one of the decisions that it might need to make. Joe Petrowski: 26:21 When it comes to the idea of a council, what is really the social contract of the council with the stakeholders, and who are the stakeholders? Gavin Wood: 26:30 The council's a fairly simple abstraction. It's based upon the Parliament in the UK. It's a set of elected representatives. In this case, I chose to use an approval voting system that's rotating, which means there isn't a single election event. Rather, there's a number of them, and over time the members sort of rotate out, so they get refreshed in a sort of round-robin file fashion, so that not all of them have to sort of go out at the same time, as is the case in the US and the UK systems. Gavin Wood: 27:07 Once they're in place, then they're free to kind of, as a whole, form much of the legislative timetable for the governance system. There can still be public votes, but they happen sort of in an interspersed fashion, interspersed where they count the votes that the council puts forward, and the council can, if they are unanimous, can sort of bias the count to a degree that means the public have to, or the coin holders rather, not just the coin holders but the lock-voting coin holders, have to vote against it with some substantial quorum before it would be rejected. Gavin Wood: 27:52 Basically, the council, if they act as a whole, as a unanimous body, get the benefit of the doubt that whatever it is they're tabling is in the interests of the blockchain, but it cannot be emphasized more that the ultimate source of authority for making these decisions rests with the coin holders, so it still means that if 50 percent of the coin holders, with full conviction, believe that the decision shouldn't happen, then they will always be able to ensure that that decision doesn't happen. They are the ultimate authority. The council are merely sort of caretakers that have the benefit of the default, but little else. Joe Petrowski: 28:36 This idea that over 51 percent can just fork the network and take it where they want, like, who are the stakeholders, and are they all part of the network? And so, a parallel I'm kind of thinking of is, like, with the recent Boeing 737 MAX crashes, the people who are in the airplane, presumably, are not part of the design process of the airplane. They have no say in this, and so with the internet affecting so many people in the world, how do you go about looking after the rights of people, or at least being mindful of the people who may not have a direct stake in the system but are affected by it in some way? Gavin Wood: 29:12 If you want to have a voice, then you need to get some stake. I absolutely refute the view that blockchains are some sort of public service, that every person on the planet have some right or some say in the decisions that should be made within the blockchain. Absolutely not, no. The blockchain is an economic vehicle. If you don't want to take part in the economic vehicle, you are free to. You are free to not take any stake in it, and you are free to ignore it. But ultimately, the economic vehicle owes you nothing. Joe Petrowski: 29:46 This is, essentially, a voluntary democracy. Gavin Wood: 29:50 It's not a democracy at all, nor should it be. Democracy is all about one head, one vote. Democracy is about all of the people within the nation having a say. Well, I'm happy for the idea of all of the people within, that hold stake in the system, having a say. Sure, weighted by the stake that they hold, but everyone who is not holding stake, they don't get to have a say. Joe Petrowski: 30:15 But holding stake is voluntary, so you get to choose if you want ... if you think you're affected by the system, then you can buy into it, then have a voice. Gavin Wood: 30:23 I'm affected by the US. I don't have a voice in their elections, but that doesn't stop the US from being democratic. Joe Petrowski: 30:30 Fair enough, whether that's right or wrong is kind of besides the point I guess. Gavin Wood: 30:35 Yep. Joe Petrowski: 30:36 When it comes to interfacing with the outside world, one of the goals of Ethereum smart contracts was to to take ambiguous, outside world data, and be able to make a very clear decision on this. What's the best way to define a program? Like, how it should be have, and then what to do when it doesn't behave as it is intended? Besides, like, the code is law approach, which doesn't really work because as a reference among implementation and code, it doesn't always behave how you think it was going to. Gavin Wood: 31:04 Sorry, could you rephrase the question? Joe Petrowski: 31:08 Yeah. So, not just in an Ethereum smart contract, but any kind of deterministic system that's difficult to mutate, you could say. If it doesn't behave the way you intend it to, what's the best way to handle the fallout when it doesn't behave that way? Gavin Wood: 31:22 I think principles are important, and the principle of the matter is that when you sign up to a system, you sign up to its rules, even if they're based upon a misunderstanding of the rules you signed up to. The misunderstanding is sort of, in principle, is you as the user's fault, not the system's fault, and not the other users of the system's fault. I think principles are important in providing a guiding light of the rough direction to go in, where we should be headed, and I think as a principle, particularly for low level systems, immutability is perfectly reasonable. Gavin Wood: 32:02 That is not to say that there are instances where principles need to be set aside in the interests of self-interest and pragmatism, and when I say self-interest, I mean the interest of the system as a whole, not the interest of any particular individual actors. Governance is there as a means to determine and codify, formally, when pragmatism and the wider interests of the system outweigh the principle of immutability, outweigh the principle of expectation. Gavin Wood: 32:39 In general, systems exist to uphold user expectations. If the governance mechanism should decide whether expectations are better upheld if the system remains immutable against, maybe, some unexpected stake transition, because of some bug in a contract for example, or whether they are best upheld, whether the expectations were such that a sufficiently large number of people in the system believed that the system would act in some way that wasn't actually the way that the code says it should act, that's the point of governance. Gavin Wood: 33:18 It's basically to work out this really soft notion of how many people expected it to do X versus how many people expected it to do Y, and the strengths of those expectations, and to balance it against the benefit of immutability, but then the costs of immutability, too. That's not something that a protocol can sort of objectively build in. Gavin Wood: 33:44 We just don't know how to quantify human expectation, and that's why we have to sort of fall back on humans and their governance mechanisms, but overall it's really one against principle versus pragmatism, and you need something like governance to work out where the balance lies. Joe Petrowski: 34:03 To go into a more technical level, we talk a lot about supply chains and sensors, oracles, so if you have sensors or actuators that are interacting with a system, a lot of these systems have floating points numbers, for example, that you are outspoken against, or calibration errors, or drift. It's pretty easy to tell if a sensor just fell off, but there are other errors that are kind of fuzzier, so how can you take these trustless, deterministic systems and interface them with the real world so that they're not just self-contained? Gavin Wood: 34:40 Well, interfacing with the real world is like input and output, so the output is relatively easy. You just say, "The blockchain wants this to happen, I will pay someone if they make it happen." Input's the harder thing, sure. You have some phenomenon in the real world that you want to record on-chain. Now, a chain doesn't want to trust any particular oracles or any particular measuring device, or measuring actor, and so we end up with something of a problem. Gavin Wood: 35:08 We want to sort of measure something, or we want to get a measurement of something, but the blockchain itself doesn't have any power to measure it itself, and it can't trust any one actor to do it for it, so what do you do? Well, you know, you have to sort of spread it out and say, "Right, we're going to use some cryptographic mechanisms to sort of select, perhaps at random, a bunch of bonded, measuring people that are going to go and measure something on demand, and then come back and tell us what it was, and then be paid on the assumption that the blockchain sort of continues to believe that they give a good answer." Gavin Wood: 35:45 Now, you know, you've got all sorts of crypto-economic, or economic mechanisms that can help in some way, shouting points being a notable one. Unfortunately, there are pretty much just as many arguments why these things may not work when the stakes are very high. I don't really have a sort of ... you know, I'm not going to come and give an amazing answer, and cover a silver bullet here. Gavin Wood: 36:14 No, I don't think it really exists. I think you just have to accept that there are mechanisms that exist, make sure that you have the right mechanism given the right sort of level of potential malicious actor or attack within the system. Joe Petrowski: 36:31 Well, that's hopefully one of the benefits of increased intractability, is that you can have increased specialization in dealing with these niche cases. Gavin Wood: 36:39 Yeah, I think that's right. The ability to create sophisticated on-chain logic will no doubt be an important factor, building the sort of things that are robust to attack. Joe Petrowski: 36:54 I want to finish up with, just for new people who would like to get into the area, what do you think are the most important skills to have, and what areas are lacking in the industry, that could use more development? Gavin Wood: 37:07 Deep thought is a really important skill. I'm not sure if it's a skill or a talent, but the ability to consume and internalize a problem, and understand it to the point that you can think about it outside of a traditional sort of trustworthy central actor, with sort of untrustworthy secondary actors asking about stuff for it, to the point that you can really imagine it as many disinterested actors all trying to interact with each other, asking them questions, and think about the sorts of rules that you can introduce that would allow the emergent effects to become what it is that you want. Gavin Wood: 37:51 That's an incredibly difficult thing to learn, but it's something that is very important in the world of this trust-free, digital, crypto-economic world. Beyond that, understanding very basic computer science things, like computational complexity, and storage complexity. Very important. Understanding what parts of an algorithm may need to be refactored or redesigned simply because they can't be implemented on-chain, because it would blow up the state too much, or it would have, potentially, an attack vector in terms of economics. Gavin Wood: 38:30 Beyond the sort of more computational stuff, I think understanding better which problems are better solved in a decentralized fashion, and which problems are perfectly fine for now solved in a centralized fashion is important, and for that it's really about understanding a little more about the underlying technology, and the costs, and the benefits to using things like trust-free systems like blockchains, and decentralized messaging. Gavin Wood: 39:04 Yeah, it's ... I mean, it's not really skills, I guess, but it's for sure any skills are useful. Right? Any skills can be turned to be used in this ecosystem, but if you're looking for sort of those really important bits of genius or talent that really push you into the next level, then I think that's what they would be. Joe Petrowski: 39:24 This is sort of a meta complexity, where you have complex systems with a lot of parts interacting with each other, and then to understand, how can you have a bunch of independent actors, whether they're people or computers, acting within various parts of multiple parts of the system, and then how that all interacts? Gavin Wood: 39:42 Yeah. It's really just ... it's almost like designing board games, actually. I did this, as well, before I came to work in this world, this ecosystem, and designing board games is all about coming up with a set of rules which at once makes it sufficiently good of a challenge, but also that people can understand sufficiently well to be able to come up with a strategy, and that's really difficult. Creating rules that are accessible, that people can understand and internalize enough to work out how to play the game. Joe Petrowski: 40:34 Thanks for coming on, Gav. Gavin Wood: 40:35 Thanks. Joe Petrowski: 41:06 Thanks for listening to the first episode of the Relay Chain podcast. Our upcoming guests include Amber Baldet and Patrick Nielsen from Clovyr, Anna Kaplan and Deirdre Connolly from the Zcash Foundation, and Jillian York from the Electronic Frontier Foundation. We'd love to keep in touch. Follow us on Twitter @RelayChain, or email podcast@parity.io. Joe Petrowski: 41:19 Our team at Parity includes some of the leading peer to peer networking developers, consensus algorithm inventors, blockchain innovators, and Rust developers. If you want to learn more about our work or want to work with us, visit our website at parity.io and sign up for our newsletter at parity.io/newsletter.