 Okay, so my name is Greg McMullen. I'm Chief Policy Officer with a company called Ascribe. We're based here in Berlin, and we've been doing digital art on the Bitcoin blockchain, but I'm here today to talk about Public Big Chain DB, which is our new product, and we'll get into that. So today we'll be talking about these four things. What's a blockchain which is more controversial than it should be? What's Big Chain DB, which is slightly less controversial? What about a public big chain DB? And if you wanted to have one, how would you go about doing it? So the first question, what's a blockchain? It's very simple, but it turns out it's quite controversial, so we'll modify it a little bit. We'll make it, what's a blockchain with an asterisk? So we'll be answering two different versions of this same question. First we'll focus on blockchain as a noun, and we'll look at the original Bitcoin blockchain, which is usually what people are referring to when they say the blockchain. It's the first, it's the most famous, it's probably the most used at this point. It's the ledger that underpins the Bitcoin currency. So at its most basic, it keeps track of who owns what Bitcoin and the network behind it makes sure that the transactions made on it are valid. So we're going to do a very simple version of this. I'm a lawyer, not a cryptographer, so if there's any crypto people who want to attack me do it after, I'm more than welcome to take the abuse because you're probably right. And the other thing to keep in mind is that for our purposes today, the exciting thing about Bitcoin isn't that it's magic internet money, it's that it's decentralized. If we wanted to have magic internet money, we could use a bank or we could use PayPal. Those are really good ways to send money to other people online. The important thing is here that we don't have to go through one of those organizations. I can trust anyone else using the network just by the virtue of the network itself. So the Bitcoin network is made up of thousands of computers around the world running the Bitcoin software. And when a new transaction is sent to the network, all of those different nodes review the transaction to make sure that it's valid. And if it is valid, it's combined with a bunch of other transactions in a block, which is added to all the previous transactions in a chain, which is where block chain comes from. So the trick is that in each block, they add a cryptographic hash of all of the previous blocks so that with each new one, you know that the previous ones are as they should be. So if someone tries to spend Bitcoin that they don't have, or if someone tries to change the history of transactions, the network will reject the invalid transaction. It drops off and is refused to be written to the chain. So there we have a very simplified version. Now for our purposes, what we really want to pay attention to is, again, not the currency, but other uses of the block chain, which are made possible by this. There's a field called opt return in every Bitcoin transaction. It's meant to mark them as invalid, but it can also store any other kind of data that you want, any string. It's a pretty small amount of data, but it's pretty useful because it means that suddenly you can use this ledger for cryptocurrency as a database that brings these really great secure features to any kind of data you want. So right away, people started using this for interesting things. There's a website called proof of existence, for example, that takes a hash of a file, writes that to a Bitcoin transaction, so you can demonstrate that that file existed at the point of time that it was registered on the block chain. At a scribe, we're doing something similar, but we're doing it with digital art. So you can upload a piece of art, you can write the hash to the block chain, you can even create unique digital editions and sell those to collectors. This was pretty exciting to the digital art world, and we started getting a lot of requests from established galleries and museums, which really led us to the big problem with the idea. We found that the limit was the Bitcoin blockchain itself. The blockchain characteristics were really exciting, but the blockchain was getting in the way. It can only handle a maximum of seven transactions per second, and in practice, it starts to choke up that two and a half or even two transactions per second. So the other issue is its size. The current block chain going back to the beginning of the network in 2009 is considered bloated if you ask the Bitcoin community, but it's only about 50 gigabytes. And in database terms, of course, that's next to nothing. So between those two things, between the low latency and the small size, we weren't able to meet these gallery's requests to register their entire collections. There was one gallery that wanted us to register 500,000 works, and it would have clogged up the Bitcoin blockchain for decades. So we took a step back and asked, what is a blockchain as an adjective? And we're starting to see lots of other blockchain projects springing up. Other cryptocurrencies like Ripple or Dogecoin, other models entirely like Ethereum or Ripple, and that starts to become an important question. So for our purposes, the characteristics that we care about are, number one, decentralized control, meaning there should be no central authority that can make arbitrary changes or decide what can and can't be done through the network. Number two is assets. It should allow people to create a token, use that token for what they want and transfer to perform actions on it to give it to other people or use it themselves. And finally, immutability. You have to be able to trust that the ledger is what it says it is, that it can't be tampered with, even by the people who control the system. So our answer to this was a product called Big Chain DB. And we're not going to do a full pitch on this today because that would be really annoying. But basically this was our answer to the question of how do we do what we want to do on a blockchain without getting all of the people really angry with us. So we decided to blockchainify an existing big data database. So we would add in those blockchain characteristics and maintain the big data speed and capacity and queryability. So for anyone wondering about this, we started with Rethink DB. And we built an add-on on top of it that we call Big Chain DB. If you want to know more about that, you can go check out our white paper or the source on GitHub. But again, that's not why we are here. So moving on, Big Chain DB uses a federated model for consensus. So rather than in the Bitcoin model, everyone being able to jump on and take part, it has predetermined nodes. And only those nodes get to vote on the transactions. So if you run your own private instance of Big Chain DB, you get to pick who those nodes are. So for example, we're doing some work with financial services companies, and they get to have their banker buddies choose who the nodes are. So either internally, different divisions or just multiple servers within their organization. But what we're really interested in is building a public version of this. So one that's operated for anyone who's working with the big chain DB, and one that we use for anyone to build their applications on top of, and for the good of the entire decentralized Internet. So what we want to talk about today is what that would look like, what we could do with it, and how we could make sure that it worked well. And we're really going to focus on making sure it works well at this stage. So I guess the first question is it works really well. It's more reliable. It's easier to use than the Internet of the mid 90s. It's fairly secure for what it's worth. So why do we need a decentralized Internet in the first place? So through the 90s, we were on this track towards decentralization. We're moving away from the mainframe model to personal computing. And we had the web and everyone had their GeoCities page. And there were lots of animated gifts. And it was a big revolution. And we even declared independence of cyberspace. And that we thought that these weary corporations and governments would just get out of the way, and everyone would live peacefully online in a society of the mind. But it didn't exactly work out like that. So centralized platforms have dominated the Internet. And these new intermediaries that have come in are just as powerful as the ones that they've replaced. And even though they talk about freedom of expression and these values, they have shareholders to keep happy. And they regularly censor content on their platforms. And sometimes it just seems ridiculous. Like earlier this year, Facebook removed photos of Copenhagen's little mermaid for showing too much skin. And their response to this was, some audiences within our global community may be sensitive to this type of content, which is clearly upsetting content. So thankfully, they have solved their problem. But we're living in this bubble of centralized content where the lowest common denominator of law of morality is being allowed to decide what we can do on our systems. Another example is the Nimors. It's about who controls our identity. In the Nimors, Facebook and Google were fighting really hard to enforce their real name policies across Google Plus and Facebook. And it had its own share of silliness, kind of like the little mermaid. Like Selman Rushdie being forced to provide his photo ID and change his name to what was listed on his passport rather than what he has gone by two millions of people around the world. So it can be silly, but it can also have really damaging effects. So those restrictions are insulting at best and discriminatory at worst and dangerous to people who have chosen to live under another name. But if you want to take part in the centralized internet, you have to live by their rules. And another problem is that the central powers make a very tempting target for other sources of power. So even if they're acting totally benevolently, having the single point of failure is a pretty major concern. We might think of authoritarian regimes targeting sites for censorship in China or Iran, but it's not just to totalitarian regimes. Just yesterday, a Brazilian judge shut down WhatsApp for three days for refusing to provide encrypted information that it itself couldn't read. The same judge had earlier detained Facebook executive for the same thing. Recently in France and outside of France, we've seen right to be forgotten or right to erasure decisions being enforced that would see Google having to wipe search results, not just in France, but across Europe. So the final reason for wanting all of this is that a decentralized stack is emerging. And for anyone who's paying attention to the decentralization world, they're familiar with Ethereum, which is decentralized processing platform running on its own blockchain, IPFS, which calls itself the decentralized web. And we see a role for a decentralized database to sort of fill in this stack and enable it. So if we assume that we need it, the next question is how do we actually make it? So we started by thinking about the structure. And we're a for profit company, which obviously doesn't lend itself well to continued spending of money on doing things that are for the good of the internet. So we knew that for it to be trusted, it would have to be run independently of our company. So we decided to establish an arms length organization, but then what kind? So we looked at the different models, we could be a different for profit company, we could be a public benefit corporation, or we could be a charitable organization. And we've decided to go with the last one, we're in the process of registering not for profit here in Germany. The other question we had was, should we be a standard organization? Or should we move into this new emerging model of decentralized autonomous organizations or DAOs? So DAOs use smart contracts to program the organization logic into a blockchain, essentially putting it out of the hands of humans. We decided that was sort of a bridge too far right now. We're inventing enough structures and building enough technology that we didn't need another layer of technology at this point. But we are building into the articles of the association, the requirement that we do decentralize at some point in the future. So eventually we will be moving away from an actual organization to a DAO, but not at this stage. So the next question is, who do we choose to be the nodes on this network? The important thing is to prevent them from collaborating, because if you have 50% agreeing to make a change, they can take over the network and do what they want with it, and you lose all of the good characteristics. So we started playing with ideas of pitting worst enemies against each other to make sure that they would never ever get along. But then we had a better plan, which was to encourage them to conspire together. So we've decided to put together a conspiracy of organizations that want to build an open, decentralized internet. So they'll always be pulling each other in this direction and working together to make sure it happens. So this was the one big criteria that we had in selecting nodes. They had to want to look out for the internet. Collusion between the caretakers isn't the only problem. If more than half of the caretakers are in one place, they're all subject to the same set of laws. So we could have a judge, like the Brazilian judge, forcing the caretakers in one country to behave in a certain way, and it's just as bad as if they were collaborating together. So we've decided to put a rule into the articles of association saying that no more than 50% can be in one country. We're calling that a jurisdictional diversity. So we're spreading things around multiple legal systems to make sure that if something goes wrong, it's going along everywhere at the same time and we have bigger problems. Money is the next problem. We have to make sure that the caretakers don't vote in their financial interest. And there are two concerns. There's direct. So milking the organization for profit, which is kind of solved by making it a not-for-profit, and indirect, which is steering the organization to enhance their own financial interest. And we've seen that happen with another decentralized network. The organization that steers the DNS, ICANN, is a great example of what happens when indirect financial interest is allowed to capture these organizations with the approval of all of these new domain names, which don't benefit the organization's voting directly, but it means that they get to sell a whole lot of new domains and make a lot of money that way. So to avoid that, we are making sure that half of the organizations have to be not-for-profit, at very least. So fewer than 50% are allowed to be for-profit organizations, which doesn't completely solve the problem, but at least points us in the right direction. And finally, how much will this cost? We had to think about how it would be self-sustaining, how it would pay for itself. And the model that we've come up with is a pay-for-use model where most users fall under a free tier, heavy users pay for transactions and for storage. And we're going to be holding a workshop with the caretakers in the next few months to work out exactly how that's going to work. So we haven't solved that problem totally yet, but stay tuned. And so now some details. What have we actually achieved here? Actually, today we might have finalized registering a not-for-profit organization, an EV in Berlin. We've established a number of caretaker nodes where more than half are not-for-profits, fewer than half are in any one country, and all of them love the internet as much as we do. And of course, we're thinking of this usage-based billing tier. So a few of the first caretakers. Ascribe a not-for-profit organization. This is going to be one. This is one of our for-profit nodes. It's also going to be our first customer. We'll be moving over to use the database as soon as possible to allow us to kind of do some of the projects that we've wanted to do but have been limited to by the scale of the Bitcoin blockchain. The second is Open Media. It's an international internet advocacy group based in Canada. And it campaigns for digital sharing, copyright for open access. And I think that makes them a pretty natural fit for us. The third is Dine.org, a free software foundry with the motto, we are free to share code and we code to share freedom. And they're behind Freecoin, which is what got us interested in them. Freecoin is a reward system that they built for the EU's descent project, which is working on direct democracy and decentralization. And finally, we have BlockStack.org. BlockStack is a not-for-profit that's developed an implementation of the DNS on top of the Bitcoin blockchain. And we think that's an important and exciting step for the decentralized stack. So overall, we're pretty excited to be working with all of these organizations. We're excited to be moving forward with this project and building an internet scale database. So any questions? I'm happy to take them now. Or you can catch up with me after. Am I allowed to take questions now? Yes. Okay. Yeah, we're organizing committee. Yeah, it seems pretty exciting. There's a lot of decentralized events popping up just on their own around the world, which I guess is fitting for decentralized projects. There's one coming up in Berlin. That's pretty exciting. There's the decentralized Web Summit in Internet Archive June 8th. You should all attend. Yeah, it seems like there's a lot of momentum picking up around it. So anyone else? Anyone want lunch? Yeah. Yes. So once we have the first batch established, we're going to turn the whole organization over to the caretakers. So the lawyer part of me is getting excited about this part. So to establish the organization, it had to be all individuals. So we've stacked it with people from Ascribe to begin with just to make signing all the papers easy. And as we add caretakers, we're going to turn over membership to them and resign. So they'll be the ones making the decisions and adding new caretakers. And we've got a process for that mapped out. It's a pretty big concern. We haven't set any strict requirements yet. We're going to be workshopping that as well. I think to begin with, no one will bother attacking it when it's just a small test net for our own purposes. But if it starts to take on a more important role, that'll be a pretty major concern. The takedown, it's up to the caretakers. So if the caretaker gets in order to take something down, they can delete it. But then the rest of the nodes will put it back up. So you believe that? No, it's a very difficult question. We've spent a lot of time thinking about that. We're thinking part of that could be done in indexing. Because once it is written, it can't be removed. It would destroy the entire database to remove an earlier entry. So we'll see, I guess. We'll see what our takedown policy is. All right. I guess we're moving on to lunch. Thanks, everyone.