 What I'd like to talk about today is the idea of scaling Bitcoin. Talk a bit about this experiment and this journey world, and some of the challenges that exist in the space. We'll continue to exist for a while. How many of you here have heard anything about the scaling debate in Bitcoin? Okay, very good. And tonight I'm going to reveal which side I'm on. Fat chance, that's not happening. What I'm going to talk about is a bit broader, this journey. I want to look at it from the perspective of some of the experiences here in Bali. Gary and others here have been building a community for quite a while in Bali. What's spectacular is that, at its peak, more than 60 retail establishments accept Bitcoin, or accept Bitcoin, for retail transactions, so you can shop for various things using Bitcoin. That's more retailers than San Francisco, which is quite astonishing, right? For a much, much smaller city like this. Now, unfortunately, probably the vast majority of those retailers no longer accept Bitcoin. There's a reason for that, and the reason is that the fees to do a Bitcoin transaction over the past year and a half have escalated quite dramatically. How many people remember the good old days from 2014? Anyone here who was part of this in 2014? Great. You could do a transaction, Bitcoin, for five cents, and things weren't easy. Now, the cost has escalated. Now it takes maybe $2 to do a Bitcoin transaction. There's a couple of reasons for that, and I want to talk about why that is, where it's going. The first part, I want to make a prediction. Think of it as a promise, a promise I can't keep, but I'll make anyway. The fees are going to come down again. We are going to see the ability to send low-fee transactions. But only if we do this the right way. There is a right way, and there's a wrong way. Like many things in engineering, the simple way, the obvious way, is the wrong way. We'll talk about why that is. I've been involved in a number of engineering projects over the years. I first got involved with internet technologies, networking, routing, communications, data communications. Back in 1988, probably even a bit before then. In 1989, I had my first connection to the internet. At the time, the only way to get connected to the internet was if you were at university. At that time, you could get a block of IP addresses and route directly on the internet with your own IP addresses. Everything was very expensive, but you could get direct routing on the base layer of the internet. Some of you may have no idea what I'm talking about. I'll throw this out to the audience. How many of you have a fully routable IP address? Actually, none of you do. Today, in order to get a block of IP addresses, you have to be an internet service provider. It is very expensive, and it is only available in some geographies that still have IP addresses left. Parts of North America, parts of Northern Europe. Asia ran out of IP addresses about six years ago. Since then, no more IP addresses are allocated to any organization. It doesn't matter how big, how important, how influential or rich an organization is, you can't get an IP address. If you want to route internet packets, your routing is translated, using a technological network address translation. Essentially, someone proxies between you and the routable internet. You can't route directly on the internet. There's a reason for that. Part of the reason is that the internet had certain capacity limitations in it. Especially when it comes to addressing. There used to be a time when thousands of people could have their own IP addresses and routes. When I first got onto the internet, I managed to get what was then known as a Class C. That's a block of 256 routable IP addresses. There's no chance in hell you can get that now. Not a chance. But then, at that time, maybe you had a university, maybe you had one IP address, maybe you had a thousand people using it. Today, I have probably, let's say, 20 devices in my house that are internet-connected. I'm sure you have the same experience. In fact, some of them you might not even remember that they're internet-connected. You set up some Wi-Fi connection on your TV two years ago. You forgot that it's internet-connected, but it actually is. Parts of your stereo, other devices in your home. A single person now has 20 devices in their home, and they're all using the internet. None of them have full access to routable space. People keep talking about the internet of things. Have you heard of this topic, the internet of things? IOT, IOT, the S in IOT stands for security. IOT is where every device in your home uses the internet, down to light bulbs. Why would you want an internet-connected light bulb? How else could you experience the wonderful experience of a hacker, turning off all your lights for 3,000 miles away? You could have an internet-connected fridge. You see that always in the advertisements for internet of things. So your fridge can order milk for you. I can order milk for myself, but that's the future we're marching into. We'll all get to experience a hacker 3,000 miles away turning everything into a rotten mess by turning it off. But what we're seeing is this explosion in the number of devices that people are using. Now you have 20 devices. Maybe a decade from now, the average home will have 500 devices that use an internet connection. You don't even know that most of them are doing it. Let's look at what happens when this starts happening to money. The same kind of progress is happening in money. In some ways, we've regressed of it. The average person who uses cash may actually do 15 or 20 transactions a day. But once you switch to using plastic, the number of transactions actually decreases. People tend to use plastic from larger amounts, and they don't just buy little things with plastic. So the number of transactions goes down. But over time, as we get more and more connected, we're likely to do more transactions. So now let's look at what that means for Bitcoin. At the moment, Bitcoin produces a 1-megabyte block every 10 minutes. There are about 3 million people trying to get their transactions into that 1-megabyte block. As a result, fees have gone up, because fees are a market-based system... for deciding which transactions are worth doing, and which transactions are not worth doing. A lot of people don't like this solution. They would rather go back to a time when Bitcoin's fees were negligible or fixed. So you could say, okay, I'm just going to do a transaction that's going to be 20 cents fixed. The problem with that is that we would end up with a system where everybody wants to get their transactions in, and now we have to make choices. So who can tell me what a good transaction is in Bitcoin, and what a bad transaction is in Bitcoin? Anyone? One that takes more than an hour. Yeah, but how can I tell whether your transaction is worth putting in versus your transaction? What's... Well, there are two ways of figuring out that question. This is really important to understand. One way to figure out this question is to have someone decide what transactions are spam, and what transactions are worthwhile. Inevitably, that someone is going to be a developer who is writing code in the software to say, you know, transactions that look like this we want, and transactions that look like this we don't want. If we go down that road, we have given an enormous amount of power to that group to decide the future of Bitcoin. Bitcoin is a system that is designed to not put people in that power, to not give that power to anyone. Have you heard the term censorship-resistant? That's one of the key characteristics of this technology, is to make it censorship-resistant. So that anyone can use it without having any control by anyone else as to whether what they're doing is worthwhile, whether it should be done or shouldn't be done, whether it should be allowed or shouldn't be allowed. So if we can't have someone decide whether a transaction is good or bad, there's only one other way to do that, and that's with a market, a marketplace for transactions, and that means having fees. That means that each of us gets to decide if our transaction is worth enough to us, that it's worth paying for it to get in now, or waiting a bit longer. And then we make choices, and I've made many of these choices myself. I'll think, you know, I want to do this transaction, but the fees are pretty high right now. I know that on the weekend, and at a time when it's nighttime in the US, there will be a lot fewer transactions in the network, it's going to be kind of quiet, and at that time I can sneak in my transaction at a much lower cost, so I'm going to wait. Lots of people making independent decisions like that, deciding whether it's worth waiting or paying more, form a marketplace that does price discovery on how important it is to get transactions in right now. And that approach gives us freedom, but it also means we can't fit any of these transactions all the time. So, we have this conundrum, this challenge. Bitcoin has a fixed one megabyte block size, and we can only fit so many transactions in one megabyte. What do we do? Anyone? Increase the block size! That's not what you said. No, I needed someone to say that, so I can continue with my rhetorical question. But since none of you were being helpful, so the simple idea is increase the block size, right? So, we make the blocks bigger. How much bigger do we need to make the blocks? Now, let me take you through a journey, so you can understand exactly how much bigger we need to make the blocks in order to get back to a situation where anyone can get any of the transactions into the network, at any time, at low fees. One megabyte, three million people, approximately using that one megabyte, to do maybe five transactions a month, mostly using it for speculation, occasionally using it for a bit of retail, and we are full. So, now let's talk about growth. When we talk about growth, one useful term is the term order of magnitude. The order of magnitude means one increase by ten, adding a zero to the end. So, three million people today. If this technology is successful, we expect more people want to use it. Let's say that in two years, we have 30 million people, one order of magnitude using it. So, to maintain exactly the same level of fees as we have today, we need a block that is ten megabytes, one order of magnitude bigger. Yes? Okay. And then five years from that, Bitcoin gets really, really successful. We need to get 300 million people to use it. Now, we need 100 megabyte blocks. And then five years from that, Bitcoin or any of these technologies becomes astonishingly successful. We need one more order of magnitude for three billion people to use this technology. And now we need a gigabyte block. Great. Now, from an engineering perspective, you can see how this goes. That gigabyte block needs to be copied to every system that's checking all of the rules in Bitcoin. That system must be able to process that block, validate all of the transactions, verify them independently in a few minutes. Not more than ten, for sure, because you have to validate the previous block before you can accept the next one. So, there are some limits as to how fast you can do that. There's that moment in engineering where your deficit starts growing faster, then you can catch up with technology. This almost goes like a joke. My daily backups take 26 hours to do. When that happens to you as an engineer, you have a problem. If your daily backups take 26 hours, that means you're not completing them in a day, and you will keep falling behind. So, if my block takes 11 minutes to validate, then I'm off the blockchain, which means that fewer people can validate independently, which means that the system becomes centralized. And this is the problem. With each one of these increases, fewer people can participate in the validation. Fewer people can participate in storing the data. Fewer people can participate in being independent actors. We go from a system that is decentralized to a system that gradually gets more and more centralized, but it gets worse. I've described to you a future where within 15 years, we're going to do three orders of magnitude increase in the number of people participating. In order to do that, we would need the three orders of magnitude increase in the size of the block. But that's if we continue doing five transactions a month at the current rate. But we want transactions to be faster and cheaper, so we need another order of magnitude increase in the block just to make it cheaper than today. So, that's a ten-gig block to hold the three billion people. But then they're going to start using it as a day-to-day currency. And what happens if they start using it like cash? Instead of five transactions a month, they want to do 50 transactions a month. And now you've got a hundred-gig block. We're now at five orders of magnitude capacity. But that's kind of pointless, because all we've done is done what cash does. And I'm much more interested in the other applications. Like, what about micropayments? What about doing very, very small payments very, very frequently? The kind of technology we don't have today. What if I want to buy an article on a blogger's website for ten cents every time I read an article and make a ten-cent payment? Well, if I start using the network like that, and I would certainly like to see applications like that, micrometering, I get into a taxi, I pay by the minute. Every minute I do a transaction, settle a small payment. But I'm also doing another transaction, because I'm paying the insurance company for one minute of insurance for the next part of my trip. Why not? Maybe I'm paying for electricity by the minute. Why not? We can do streaming money. We can really start doing things that have never been done before. But that will require an order of magnitude increase in order to do small transactions, and another order of magnitude increase in order to do micro-transactions. But why stop there? Because I really think we should be doing nano-transactions. And now we're talking about eight orders of magnitude. So now we need to ship a petabyte block every ten minutes. And not just chip it, process it, validate it. We can't win this game. There's no way that the system will scale. The system doesn't scale. And so immediately people who look at this go, we can no longer fulfill the fundamental promise that I certainly talked about, which is to serve the underbanked, to serve the unbanked, to serve the underserved. Is that out the window? Is that no longer part of the plan? Have we given up on the idea of bringing in really people who have no access to financial systems today, and giving them access to something that is a viable system of payment for a person who earns $1 to $2 a day? Not with the current transaction fees. Something has to change. And here's the issue. The issue is we have to decide what it means to serve the underserved. Because we can serve the underbanked. We can serve the underserved. We can serve them off-chain. We can serve them on private systems. We can serve them by giving them access to intermediaries. And you know where this goes. We built a system that holds their money for them. I'll just pick a random word to describe that. Let's call it a bank. And here we are again, reinventing the exact same structure of the past. To debank all of us, and bank the unbanked, or debank the unbanked with this new technology, all we need really is a bank. We can serve their financial needs only by sacrificing their privacy, security, autonomy, and liberty. That is the deal. So how do we reconcile these two things? Because this is when we start making the important choices for the future. And let me tell you what, on the internet we made the wrong choices. And we're now paying the price. On the internet, the base layer didn't have enough encryption, it didn't have enough privacy, it didn't have enough anonymity. We said, let's build those on the next layer. Well, some people did. I certainly did. I was one of those crazy little cypherpunks going, Encrypt everything in 1990. The NSA are watching us. There's something deliciously ironic about being the paranoid cook. For 20 years, until someone whistleblows and turns you into the rugged realist overnight. Turns out, they were surveillance, and they are. And so what happened on the internet was that we didn't put enough encryption, we didn't put enough privacy, we didn't put enough security in the base layer. And when that layer got concentrated, control was a lot easier. How many IP backbone providers are there in the world? Anybody have any idea? Okay, five. Five companies were out more than 90% of all internet traffic worldwide. Five companies, three of them are American. All of them are 100% owned by all of the intelligence agencies. Not just one. They're like jockeying for position. The NSA is listening, and then the Germans are listening to the thing that the NSA just listened to, and the French are speaking, backing on the back, going, hey, can we listen to, and the Australians are coming in, and the New Zealanders are coming in. And they've got this fantastic deal whereby every single one of the intelligence agencies is prohibited from surveilling their own people. So they outsource it to their partner agency in New Zealand to surveil Americans. Will the Americans surveil the New Zealand people, and everything is legal? Legal. And that's how the internet ended up serving all of us, while undermining our autonomy, privacy, security, and liberty. And we can't fix that on layer two. We can't fix that with TOR, VPN, anonymizers, and remixers. Because if layer one is not only compromised, it's much harder to fix this at layer two. Much, much harder. When we try, we all do. I do. You should. But the problem is that in most market systems, privacy is an externality. Every one of us makes a small decision about how we're going to spend our money for connectivity, for data, for access to social media. And for every dollar we save when we make that choice, a little bit of privacy dies. But we don't notice, because that happens in a very diffuse way. We make a choice, and that choice is ours. But the impact of that choice when a little bit of privacy dies is shared by all of society. In market systems, that's called an externality. And so, over the years, democracy starts getting sick. It gets a cold, and then it gets the flu, and then it gets the bubonic plague, and then it catches a bad case of the Trump. And suddenly, institutions of creativity and dissent, and thinking out of the box, and being different, and not agreeing, and being annoying, and being young, and having opinions, and saying, no, become criminal acts. And that privacy gradually means that our democracies are dying bit by bit. But it's not on your Comcast bill. It's not on your Internet Service Provider bill. It doesn't say $49.99, but $10 discount, the death of democracy discount, $10 off. It doesn't say that. You don't know that. So you make the choice, and gradually things change. And you look back 10 years later and say, how did we get here? How did our institutions die? We're not going to make the same mistake, I hope, for this system. So we have a choice. How do we scale, while preserving the principles of decentralization, independence, privacy, security, autonomy, liberty, in the base layer of these cryptocurrencies? The only way to do that is to move to additional layers. We're going to move off-chain. Many of our transactions will not be settled on the base Bitcoin chain, or on the base Ethereum chain. They will happen in additional networks. In engineering terms, you may have heard the term side-chain, or lightning network, or drive-chain. These are all things that are going to allow us to do thousands, and then tens of thousands, and then millions, and then tens of millions, and then hundreds of billions of transactions a second, without all of them getting settled on the private chain. But only if we do it right, only if we preserve the principles of privacy, security, liberty, autonomy on the primary chain, we have to get it right. That means making the primary chain slightly more expensive, slightly less scalable. It's a very big mistake to think that the underserved and the underbanked are not being served, are not being banked, are not allowed to participate in the financial system of the world. Because of a problem of scale, it has never been a problem of scale. The financial systems that deliver financial services, the traditional systems, wire transfers, digital payments, plastic cards, debit cards, credit cards, banking, they can scale, because they have already sacrificed all of the principles in order to give scale. They can do very cheap transactions. So why do they leave 4 billion people out of the system? Because of this devil's deal, this coalition with the surveillance system that says that, in order to give these services, every transaction has to have a known identity, every participant has to have identity cards and identity papers, which 60% of the world doesn't have. Every endpoint has to be identified and tracked, and it can't go across borders, and it has to be within jurisdiction, and it has to follow these complex laws. You add all of these things together, and you have the cost of politics. The cost of politics is really simple when it comes to banking. In order to create the illusion of security for a bourgeois Western elite to pretend that somehow this stops crime or terrorism, or the financing of organized crime, to pretend, because of course it doesn't do any of those things, as we have seen again and again, we are willing to sacrifice 4 billion people to poverty and say they do not get to participate, because they have not yet proven that they are not criminals, not terrorists, not other. Bitcoin and the systems like it present a completely different bargain. Anyone can access, open to everyone. You do not have to prove anything, you do not even have to prove you are human, because you don't have to be human, you can be software. Anyone can have a bank account, as long as they give up on the bank part, and we can serve the underserved. But in order to do that, we have to do it the right way. By understanding that serving the underserved doesn't mean simply delivering fast, cheap transactions. The best way to deliver fast, cheap transactions is the fascism of the central database. It is central banking on a global scale. It is totalitarian surveillance systems of a cashless society where no one has voice, no one has recourse. And the moment you go to the wrong protest, say the wrong thing, read the wrong book, they turn you off as a participant of society. This is happening today, and it will get worse and worse and worse, unless we change the bargain. And changing that bargain isn't going to be easy, and it will mean expensive fees. And sometimes these expensive fees hide something else. Gary told me that a lot of the retailers who stopped taking retail transactions in Bali, they're not that upset. The Bitcoin they earned quadrupled in value. It was no longer suitable as medium of exchange, but some need made one hell of a store of value, certainly compared to the Rupiah. And so that is perhaps the bargain we see there, and it's going to have been flow. Sometimes it's going to be cheaper, sometimes it's going to be more expensive. And if this is successful, we will be able to serve billions of people with billions of transactions a day on a global network with privacy, with liberty, with autonomy. Thank you.