 Rhaid i'n meddwl, Paul. So, Louise asked me if I would come along and explain a little bit about what the blockchain is and sort of how it works and why anyone should care about it and what the implications might be and so on. And I said, fine, that's no problem. I have a standard 14-hour presentation about this. We probably want to have some sort of interval. She said, no, 13 minutes. So this is going to be quite a challenge. I'm quite looking forward to it, but it means we're going to have to move at a little bit of a clip here and there. So let's go. So you don't need to know any more about me. OK. So I'm going to talk about, oh, these are the old slides. OK, all right. So I'm going to talk about, well, slides that you did yesterday, I feel are a little stale. These are yesterday's slides. I had some slides I did this morning, which... So I need to talk to you a little bit about transactions, first of all, in order to get you in the right frame of mind. So we want to record transactions somewhere. It doesn't have to be money, like just transactions of any kind. So we want to record these transactions somewhere. And in the old days, when computers and communications were really expensive, we had to record all those in one place because that's all we could afford. And we had to... It was like mainframes and whatever. And then when computers got a little bit cheaper, we could store them in a few different places, and we could have backups and masters and slaves. And then computers got a little bit cheaper and we could spread them around even more so we could speed them up and get access. And now you picture in your head, or at least I do, there's the bank and the bank's got a big computer and all of your stuff. But of course that's not really true. The bank has lots of computers that are all over the place. In case there's an earthquake somewhere or whatever or someone pulls the plug out while they're hoovering. But we've gone over a bit of a cusp because the cost of computers and communications has fallen to the point where actually we now no longer have to have those structures in the overheads that go with them because everybody can have a copy of everything. And we can do an instantaneous calculation, which I just thought of, so I don't know what the answer is going to be. We can do an instantaneous calculation just to show you what I mean by that. So let's imagine for sake of argument there are, I don't know, 50 million bank accounts in Britain. And let's say on average, everyone of those accounts did, I don't know, two transactions yesterday. So that's 100 million accounts. And each of those transactions was, I don't know, about 1,000 bytes or something like that. So I can't do that in my head, but it's like 10 to the 5... Sorry, 100 million is 10 to the 8 transactions. And 1,000 bytes is 10 to the 11 bytes. So that's 10 to the 8. So basically you could fit it all on my phone is my point. So you could take all of the transactions that happened in all of the bank accounts in Britain yesterday. You could fit them on my phone. So everybody can have a copy of everything. This is the starting point. Now, if everybody can have a copy of everything, it's very easy to say that. So I say, okay, so let's imagine that everybody in this room has a copy of all of yesterday's transactions on their phone. So then you have to think about, well, how's that going to work? Because first of all, somehow you've actually got to get the copies of your transactions onto the phone. So that's the communications layer in my model. So somehow I'm going to get all of the transactions to everybody in the room, and I'm going to get them to all of your phones. It's not important to the talk. We have to say it's bank transactions for the moment. Now, it's very easy to say we'll get copies of all the transactions onto all of your phones, but it doesn't work like that in real life because some of you, you've got your phones turned off. Some of you, your phones have got bugs in them. Some of them have got malware. Some of you are agents of a foreign power hell bent on destabilising the system. Some of you are hackers. So we have everybody has a copy of everything. Actually, that wouldn't really be true, right? Not all of your phones would have exactly the same copy of everything, either because of mistakes, errors, or because you're trying to worm your own transactions into the system and defraud somebody. So we need some mechanism. Like when we come to do something, how do we know what's like, you know, Paul's phone says I've only got 10 quid, but Sam's phone says I've got 15 quids. How do we know which version of the truth is correct? And so for that, we need some kind of consensus protocol. We have to find consensus so that we've got a sort of shared world view. And there's different ways we could do that. So for example, you could all look at your phones and just vote, right? And most people think I've got 15 quid and less people think I've got 10 quids so therefore I've got 15 quid. Then we all agree I've got 15 quid we can carry on. Or we could have more complicated mechanisms for doing that. We could do multi-round, like we could vote and then you could see everybody else voted, then you could vote again. Paul's vote counts more than mine because he's more important and so on and so forth. There's lots of different ways we can do this. Don't worry about how it's done for the moment. We can find consensus between all of these copies. Now one way of looking at that, so now you've all got a copy of everything on all of your phones, right? And we've got a mechanism for establishing consensus. It's like we have a new kind of computer that we never had before, a consensus computer. We've got a weird kind of computer which is actually all of your phones and it has consensus so we can do operations on it. Just as the interesting thing about a general purpose computer is you can run general purpose programs on it, that's why computers are fun and interesting because you can change the software and make them do different things, the same is true of the consensus computer. On the consensus computer we can run a new kind of application. These consensus applications which didn't exist before. We have a new kind of computer and we run new kinds of applications on it. You just saw the first four. This is the second four of my 4x4x4 model. This is the second four and they're easy to remember because they're the communications layer, the content layer, the consensus layer, and the contract layer. Contract is a stupid word for this layer and you see everywhere on the blockchain smart contracts. I don't want to labour the point but they're not smart and they're certainly not contracts. Choosing the name contract has actually set back the thinking in this space quite a long way, so I don't understand marketing. If you want management to remember things, CCCC is a much better way of doing it. We had the first four. This is the second four. Now you can make different choices each of these layers. You can have a million different ways of implementing those communications protocol and you can have a million different ways of implementing the consensus protocol and so that would be a million times a million, like a gazillion different ways of having these shared ledges work but obviously because I'm also an expert in management and leadership and if you say to people there's a gazillion different ways of doing it, you don't get very far, that wouldn't be the McKinsey way of doing it. The McKinsey way of doing it would be it will cost you a gazillion pounds for me to think about it. Hey, we've got this interesting new technology and there's a gazillion different ways of doing them really work. I normally find especially with senior management we need to think of like somewhere between three and five different ways, four I think will be a good in this case. I can get to the four in a very logical way. Remember the communications layer and the consensus layer. Forget about the content layer because it's not important. We ask a question at each of those layers. I ask you a simple question at the communications layer. Remember what I was talking about, everyone in this room has all the transactions. The first question is does everybody get to have all the transactions? If the answer is yes, then that's a public ledger a public record of all the transactions. If the answer is no, it's a private. This is not too technical for you, Paul, are we? If everybody can see it, was public not a safe word? Nobody told me about this. I was supposed to get a list, Louis and it never happened, sorry. Public! If anyone can use it, it's a public ledger and if only some people can use it, it's a private ledger. Are we okay at level one? That only gives us two different kinds of ledgers. That's not really enough for senior management, so we need to push it down a little bit further. Now I can ask another question which depends not on the communications layer but on the consensus layer. I can say in the case of the private ledger, so only some people can see all these transactions. I want to ask another question about the consensus. In other words, who establishes the consensus? Does everybody in the group who's allowed to see the transactions do they all participate in the consensus-forming process? If the answer to that is yes, then it's called a permissioned ledger. You need permission to use the ledger. Once you're using the ledger, you'll take part in the consensus-forming process. If the answer is no, if only some of you get to take part in the consensus-forming process, then that's a double permissioned ledger because you need permission to use the ledger and you need permission to take part in the consensus-forming process. In private ledgers, we have two kinds, permissioned and double permissioned. On the other side, when we talk about the public ledger, we ask a slightly different question about the consensus-forming protocol which is related to the tragedy of the commons. To form consensus requires effort. You've got to do something for it. If it's a public ledger and anybody can use it, why would you bother putting any effort into maintaining the consensus because if I'm a criminal, it's in my interest to help form consensus because I'm going to put my own transactions into that ledger. If I'm not a criminal, why would I bother? Let somebody else do it. Here it's not who takes part in it. It's public. Anybody can take part in it. But why would they take part in it? There are two answers to that question. There's the intrinsic and extrinsic answer. I can give people incentives that are not part of the ledger. That's the extrinsic solution. I can give people an incentive which is not part of the ledger. The ledger would be, for example, money. I can give you money to take part in that consensus-forming process. Or it might be a condition of whatever business we're going to build on top of it. You could see in it, I don't know, if you're thinking about storing something on your blockchain, I don't know, cars or refrigerators or something, then it might be a condition of being a car salesman or a refrigerator salesman that you have to take part in it. If there are incentives that are part of the ledger itself, that takes you into a slightly different space. In the first example, if I'm going to give you incentives for taking part in the consensus-forming process, I can also withhold those incentives. In other words, if you form a consensus that I don't like, I can not pay you anymore. I can block you from doing it. But if the rewards are part of the ledger itself, I can't do that. That gives us two kinds of public ledger. Everybody can use the ledger and everybody can take part in the consensus-forming process, but I can withhold the incentives. That's permissionless, because you don't need permission to use the ledger, but you do need permission to form the consensus. Alternatively, if it's a public ledger and anybody can use it and nobody can stop you from being part of the consensus-forming process, that's a double permissionless ledger. That gives us four kinds. Everybody knows the obvious example of the double permissionless ledger is somebody, come on. Bitcoin. The obvious example of the double permissionless ledger I can't stop you from using it and I can't withhold consensus is Bitcoin. I should say, by the way, I'm not used to talking to people like you, I'm used to talking to banks, and if I'm talking to banks, one of the other interesting aspects of this model, quite a simple model, absolutely nobody has a requirement for a double permissionless ledger, which is why I have a slight scepticism about Bitcoin's long-term. In other words, a double permissionless ledger, nobody can stop me from transferring this value from one person to another. That sounds good if you're a 20-year-old MIT computer science graduate. This is the mom can't stop me from getting a playstation view of the world. It's that teenage pseudo-libertarianism. That sounds good, but if you think about the real world where grandma's going to get an email and press on the wrong button on her email and suddenly a house belongs to someone in Minsk, this doesn't sound quite so good. It sounds good if you're a computer science graduate. It's not good if you're in the real world. We don't really want that sort of thing. See, 4x4x4, so you've got the four layers that, by asking questions about those four layers, that gives us four different kinds of ledgers to think about. Now, what's the narrative? What are we going to do? This is I slightly amended the quote by adding a hashtag. I'm a totem of modernity in this respect. Oma Kaian thought of the first blockchain, but he didn't have to mark it properly. I feel like that's where I step in. So what's the narrative around this blockchain? Now we have these four kinds of ledgers and we want certain characteristics of those ledgers and one set of characteristics might be that we want ledgers that can't be subverted, can't be changed, blah, blah, blah, and we might want to use a blockchain to implement those. There are lots of kind of shared ledgers, some kinds of shared ledgers we might want to implement with a blockchain. Okay, but there are also different kinds of blockchains, so what is the blockchain, which is what people talk about when doing the blockchain, that's just one kind of blockchain that we might want to implement just one kind of shared ledger. Okay, and there are different kinds of these, I mean I've put three examples out and don't worry about what the examples are, I'm just illustrating the point there are different kinds of blockchains that we use for different things. It's a very common mistake thinking in this space to imagine that the particular blockchain that's used for the particular implementation of Bitcoin is the generic solution in the blockchain space. It isn't, it's a very special solution to a very specific set of requirements which none of you have. Okay, now what are we actually going to use those ledgers for? Again, don't worry about how it works, there's clever cryptography that makes it work. We have a ledger that's shared blah blah blah, don't worry about how it works, we can make it work. What are we going to do with it? I think that's a much more interesting question. And what we're going to do with it is we're going to make some trust as transactions systems so perfect that no one will need to be good. And my suggestion if you're looking for something radical to do in this space, my suggestion is to focus on the transparency aspects of the blockchain, not any of the other aspects. So it has certain characteristics right? All the transactions are on all the phones in this room, that makes it very resilient, you know if I'm an attacker I've got to take over most of the phones blah blah but can we know all of those things? But a more interesting characteristic is we now all have all of the transactions. Now that doesn't necessarily mean we can all read all of the transactions because the transaction might be encrypted in certain ways and in the general case that would be true. I want you to be able to look at the transactions and see that my bank is solvent but I don't necessarily want you to know what each transaction is. That's a property that I call translucency, we want translucent transactions. We need to be able to see that those transactions are correct without necessarily seeing what they are and if that sounds slightly counterintuitive just take it from me, we can do the math we can make that work, the cryptography is there to do that. So I want to focus on transparency in order to build my narrative. This is a very favourite, I'm a naturally very conservative person, this is a very favourite picture of mine. This comes from the 1930s this is a picture from popular mechanics magazine, I think in 1936. It's drawn by a very famous New York architect called Francis Kealy. He's very famous because he was one of the first people to design actually as it happens in airport but he's famous for other things as well. He was challenged to come up with an interesting solution to a very 1930s financial services crisis and depression in America which was bank robbery. It was an epidemic and it was terrible and people were getting hurt and blah blah blah. It was a slightly different problem in those days because in those days the people that robbed the banks didn't work for them so there was like guns involved and it was... There was like a different dynamic to the whole thing in those days but nevertheless it was a serious problem and so people came up with all, you know let's give the tellers machine guns and let's have steel things and bars and Kealy came up with a different solution and he said why don't we build banks out of glass? That's a glass bank and he said if you try and rob a glass bank it doesn't work because everybody that's walking past can see what's going on inside, they can call the cops, right? He said you can't rob a glass bank even if you work for it because everybody can see what's going on and I want you to focus on that I'm not suggesting you actually build them out of glass I'm just making very clear this is a metaphorical glass bank but you get the thing what I'm talking about so he said if you build banks out of glass you get a more honest system and I think if you want to locate what to do with the blockchain if you want to take the product of the new technology and build something with it it doesn't matter if we can send the money around a little bit cheaper, a little bit quicker that's not going to make any difference but the fact we have this new kind of transparency this translucency really does and so if you think that's a silly example this is a police station in Tbilisi in Georgia the Georgian government is having anti-corruption crackdown at the moment and as one part of that crackdown they've started building the police stations out of glass so everybody that can walk past can see what's going on inside they've actually made a physical version of that and the architects that build these glass buildings have a very lovely word for it which I'd like to appropriate they call it ambient accountability we build systems that have ambient accountability that's what's interesting about the blockchain in your context we build a system with ambient accountability we dissolve the traditional boundaries around auditing and inspection we make the environment itself responsible for maintaining the trust in the system so don't worry about how the blockchain works people like me can make the cryptography work but if you can think of ways to use ambient accountability you can create entirely new markets entirely new kinds of financial services thanks very much for listening so I wasn't sure if you people used Twitter so I put a few suggested tweets up just to help ease you along into the whole kind of Twitter thing just as a sort of public service thank you for listening