Reimagine global payments with the creators of Interledger

Their conversation covers a variety of topics and reflection. We dive into their early days and the challenges they've faced in creating a universal payment system and a role of ILP in enhancing efficiency. They dive into smart contracts, decentralization, and a need for a universal payment network that connects different ledgers. The conversation highlights the successes and the mistakes made along the way. And of course, the ongoing work to improve and integrate technologies.

We hope you enjoy another episode of the Future Money Podcast.

As a freelancer, I knew that there was a lot of friction there. Working with people in other countries was always very difficult. Just getting them paid and getting the money there. So Bitcoin didn't fully live up to that by itself. I think it's a great foundation. But I ended up joining Ripple, uh, working on XOP Ledger, um, working on other blockchain technologies as the chief technology officer at Ripple.

And as part of that also created a research team, which Evan was a part of and he'll tell you his perspective on that. And yeah, during that time researching things and trying to figure out like what are some of the missing ingredients to take. Blockchain technology and other existing financial technology and kind of bring it up to the, to the level that we'd like to see where everything feels really frictionless, um, that led to the development of Intellidger.

And so nowadays I am the chairperson of the Intellidger foundation and I kind of work full time on open source code. I work on a project called Dasi, which we'll probably get into later. And I'm just trying to get Interledger out there into the world.

This is like a 14 person company talking about connecting the global financial system more like that seems nuts, but also very interesting. And that was kind of how. How we started working together.

The math just doesn't math and, you know, you're just doing too much work for each transaction. And so, you know, in the traditional finance world, a traditional payments world, the way that you often deal with that scalability issue is with netting. So, like, if you make a credit card transaction, there isn't like an international bank wire that one to one corresponds to your credit card transaction, right?

Like, The way that works is that the banks will add up all the transactions and then send one big transfer to, to net it all up. And so I always thought it would be something like that happening for these newer technologies as well. I just didn't know exactly what it would look like.

And I remember at some point, Stefan saying, I don't know if it's going to be possible to get everyone on one ledger because like there was the like Bitcoin sort of started off the whole blockchain thing. But then a lot of other people came out with with different blockchains with slightly different features.

And I think like one of the things that I always give a lot of credit to Stefan for was like the company's whole mission was like, let's get, let's connect the entire world's financial system. Let's connect everybody to this one ledger. And I think the way that I remember Stefan looking at it was like, Oh, what if that's not possible? Is there a way to achieve this vision in a different way? Uh, and I think that was the current, that was one of the kernels that kind of kicked off.

And I basically just like took her into a meeting room and I was like, so this is going to sound a little weird because you know, our whole thing is that we want everybody to connect through the XRP ledger, but, you What if there were multiple ledgers and we had to think about how to connect them all together and like, what if we changed our view and, and the XRP ledger is just one of many, many ledgers.

And that seemed initially very like counter to the company's goals and took us a long time to kind of figure out how it all fits together. But I think looking back, you know, even with the original vision, you still have to deal with legacy ledgers. You still have to deal with bank internal ledgers and companies, internal ledgers. So the world has many, many ledgers and you need a way to get across them.

So that, you know, that was sort of inevitable, but yeah, it was very scary at the beginning.

Of sending money across them so that you and I could be on different payment networks and still send money to one another I think that's the simplest way that I can put it It's also worth adding that we in thinking about the design for how to actually do that We took a lot of inspiration from the design of the internet, which is also why the names are somewhat similar But the internet is also about routing Packets of data, not money across independent networks.

So I'm on a different internet service provider than you guys are. And that's fine because the internet just makes the connection happen in the background. And that's kind of the hope with interledger is that we could be on completely different payment providers and still be able to send money between one another.

Right. Which is, I would say just making money move very efficiently and, you know, getting to a point where, you know, communication has become so ingrained and effortless, you know, you don't think about like, Oh, you know, we're gonna have to write a letter and I'm gonna have to take it to the post office. Like you just type whatever you want to send to somebody and you send it. Within seconds, they can read it.

And we haven't quite reached that level yet with, with payments, even though they're also just bits and bytes. And so IntelliJ is essentially trying to, to catch payments up to the same level that the data is at. So that's, that's one side of it. And then when you talk about the technology, I think Evan kind of already covered, like it's very much inspired by some of the key ideas behind the internet.

One of them I would highlight is this idea of just it being an abstraction layer, it's just an API. Like you're, you are. Building applications for interledger. And you don't know if they're eventually going to run on some blockchain or they're going to run on some bank ledger, they're going to run on some other system. It doesn't matter. You're programming your application for interledger, and that means it's going to run everywhere.

So it's kind of like, um, connecting all the different solutions together. And then I think the other thing, which is maybe getting a little too into the weeds, uh, at this stage of the conversation, but I'll mention it briefly, which is the end to end principle. So the internet.

Kind of was competing it at the time when it was being developed with other, lots of other network interoperability standards, because people realize, okay, we have all these networks, but you know, kind of would be nice if they were all connected together. So there were a bunch of different efforts along those lines, but I think one thing that really helped the internet win and become the standard that we all use today was that it was able to be much more simple.

And the reason it was able to be much more simple is because. It realized that there are a lot of features that you can provide on the network level, like reliability, checksums, you know, also the things like that. But if you actually skip all of those and you make a deliberately unreliable core network, that core networks ends up being very simple. And then what you do is you treat all those. Issues or you handle all those issues at the end points.

So when I send a packet over the internet, I don't know if it's going to get there. Maybe not. Maybe it will, maybe it won't, maybe it'll take a while. Maybe it'll get there tomorrow. Like, I don't know. But I will handle those different failure modes and you kind of have to do that anyway, because whatever reliability mechanism you implement on the network layer could fail.

And so you don't really lose anything by doing it this way, but you gain a lot of simplicity or like you lose a lot of complexity in that sense. Right. So we tried to take that same architectural approach with Intellidger. And if anything, you know, I mean, first of all, credit to Evan, because he was the first one to kind of think in that direction. And that ended up being, I would say the key breakthrough is like. Really using the internet as the foundation.

Um, for the design and we've been surprised again and again, how well it actually translates over. Like there've been so many things where I was like, surely payments are different. Money's different in this way or that way. And it ended up being like, no, actually we can just do what the internet does, you know? So it was very interesting.

I think a lot of internet traffic, I don't know what the numbers are, but, uh, probably the majority of internet traffic is still using IPv4 today. Okay. But there was sort of one big issue with IPv4 was that it was meant as a sort of a testing version. It wasn't really meant to be the thing that then takes off because it had a limited address space. When I say limited, it's like 4 billion, a little over 4 billion addresses. So it's, it's not that limited.

But when you're trying to connect the whole world, like it does start to become an issue. And with the number of devices that are trying to connect today, it's definitely become a bottleneck. And so that was the main driving force. But of course, when you're doing a revision to the core protocols, you try to make some other improvements. You cut some stuff you didn't end up needing. You add some stuff that was missing. And so there's a new version of the internet protocol called IPV6.

And people are trying to roll it out and have been for many decades, several decades. And. We're sort of in a similar spot, I think, where like, you know, there is IPV4. It works great. It has all the things you need, but we're always thinking about like, what are some tweaks we could do here? Some, some things we could clean up. Um, I think one thing that is sort of a mistake that, that I pushed for at the time was the date format in the Intellidger packet is sort of a textual format.

I'm not going to go through all the reasons why I thought that was a good idea, but I think like, Today, even when I'm writing my own implementations of it, I'm like, I wish it just, just a number, just like a lot of network protocols do just use a, an integer, um, so kind of wish it was that there is a congestion control mechanism in the internet protocol called explicit congestion notification.

CN again, it could get very technical here, but basically it allows the network to detect when congestion is about to happen without actually having to drop any packets. So it increases efficiency. The internet was able to implement it because they had some bits available that they could sneak in and make these the ECN bits. We don't really have any spot for that in the intellectual protocol format. So yeah, we might make some tweaks and stuff.

That's actually one of the conversations I'm very excited to have at the summit in Cape town later this year.

And we had a couple of different protocols that were part of what we called ILP at the time. We had a couple more fields in the packet. We debated, do we need a packet format? I was the proponent of not until convinced that it was a better idea to have a specific format and we like really poured over every aspect of it to try to see, is there anything we can rip out of here?

And when we got to the What we considered at the time the final version, or I think still sort of do, or I still do, let's say, uh, we had that feeling of there's nothing left to take out for the thing that Stefan mentioned, I would be curious whether, like, that extra bit would be possible to add in a non breaking way so that, like, maybe there's a version 4. 1, but Ideally, not a five or not a five for a long time, because I think there's a lot of value in maintaining interoperability

and not making breaking changes.

And it defines what over there means, but it doesn't tell you anything about how to get over there, how fast you need to get it there. How like it's just like a way of saying over there. And then on top of that, you build all of the reliability stuff that Stefan had mentioned.

All of the web and all those technologies are higher level and the whole idea with the end to end principle that Stefan mentioned is that by trying to simplify the core as much as possible, you make it so that we don't really have to agree on that much across the entire thing, and that allows it to evolve on the higher levels and the lower levels. So the technologies that the Internet was built on have changed Many, many times. And there's been a lot of innovation on the networking layer.

And there's also been new protocols that have been built on top of it. And that's one of the things that's, that's really, really cool. Interledger tries to take that same approach so we can have new, new ledger technologies that it's built on and then new application protocols that are built on top of it. But the idea is that the interledger protocol should stay the same because it's just trying to define.

Can you get this money over there and similar to the Internet, it's defining how you address things and how you say over there, but not saying how you get there or anything like that.

And, um, you know, Evan was the one who developed a lot of stream and designed it and implemented it. And I hadn't even really looked at the implementation back then. And so over the last couple of months, when I was re implementing it, Just to kind of learn it better, I really got a much deeper appreciation for all the thought and design that went into it. So my, my answer would be the stream protocol. I love it. It's really cool.

Transport layer protocol for interledger can we take inspiration not from the original version of the Internet's transport protocol, but also from what they learned in the subsequent decades. And so a lot of the time spent developing stream was basically like I had the RFC for a quick open on one. In one tab and then a parallel one was like writing, okay, that seems relevant for us too. Great. Let's copy that idea. Oh, great, great, great.

And that was, that was a great jumping off point to build that protocol. So most of that credit belongs to the quick authors.

Right. And so I think that, you know, I, I will say that I'm personally made a lot of mistakes in terms of like the strategies to pursue and how to drive adoption and some of the decisions along the way, but thankfully those are all, you know, transitory, right? Like you, you make a mistake in the strategy for adoption, you learn from it, you try a different approach. The parts I'm much more happy with is this, how the actual protocols are designed and everything.

But in the vein of like how, what's a better strategy, the best success that we've had with Intellidger were probably in 2019 when we had sort of a community network built around Codius and Intellidger and people could run their own Intellidger nodes and things like that. And I remember that time as being very dynamic. Like this was the time where, you know, somebody made a dashboard for seeing the status of the Codius network that we had no idea this was coming.

We don't know, I think we still don't know who built it or just, I don't know. Um, and it just sort of popped up and it was amazing. And I really regret that at the time we sort of said like, okay, we have to move towards like more regulated institutions and working with these like, you know, traditional financial institutions. And so the community network kind of needs to take a step back and we stopped supporting it and we kind of actively discouraged people from doing that.

And I think it was a huge mistake because for a community to grow and thrive, like people need to be able to actually do something. Um, and participate and, and so on. And so I wanted to bring some of that spirit back to the project. And, you know, we are living in these amazing times where there are all these blockchains and you can just, you know, write some code and it'll connect up and you generate a private key. And so suddenly you can send money around.

And so I was very interested in bringing that feeling back. Dossi is a peer to peer implementation of Intellecture. So rather than. It's being meant to be run by centralized organizations. It's sort of something that you just download, you run it, and it does stuff. It does intellectual stuff. Um, it speaks all the intellectual protocols.

From the bottom of the stack all the way up to stream, uh, and even open payments, which we haven't mentioned yet, but that's sort of a, how to negotiate a payment, like how much should I pay that sort of thing. And, um, it's my hope that, uh, people try it out and play around with it and learn about intellectual that way and hopefully drive the adoption forward.

It is designed to challenge students to develop long form audio productions, including podcasts, audio books, film scores, and more. Through this project, students created interstitials about Interledger for our Future Money podcast. Here is a sample from Shamari Hall, a 2023 graduate of Howard University. Stay tuned.

org slash summit to learn more and explore how digital finance Can benefit everyone, every way.

So Bitcoin had this sort of very minimal, uh, stack based scripting language built in. So we were thinking about like, what would it look like if you built like a general purpose programming language into a blockchain? And, you know, somebody was crashing on my couch at the time was Vitalik, so he clearly got somewhat inspired by that work.

And so, you know, that sort of evolved over time and I think that got to a point where I was like, okay, why are we even trying to build all the logic into, into the blockchain itself anyway? Like the, a blockchain is this sort of thing where, as I mentioned at the beginning. You're running everything many, many times. And so it's very computationally expensive already. So why are you doing all of your logic on that layer?

The way we do it in most software architecture is we have a database and then we have our business logic separate from that, you know, you have a separate layer for that. And so, you know, we started to think about like, okay, you know, this was Evan and I back in the day as well, you know, thinking about like, okay, so we have the blockchain, which is kind of our data layer. Well, what would a decentralized business logic layer look like? And that became the Codeus project.

And the basic idea is just, okay, you have these stateless systems, so they are not databases and all they do is they run your code, but you can have many of them. And they then might make transactions against the blockchain. They might query some traditional API, et cetera. It's, it's actually kind of a pretty simple idea. You just have this like group of nodes that does stuff for you. And that way you implement smart contracts in a more flexible way.

But the problem we ran into very quickly was, okay. So we have this very neutral smart contracts layer, you know, the smart contracts layer, that's not tied to any particular blockchain. You could use it with any blockchain. You could use it across blockchains. You could use it across non blockchain systems, but you kind of have to pay these hosts that are running the code for you and. You could pay them with a credit card or something like that. But that seems very clunky.

Like you're trying to pay a hundred different hosts with a hundred different credit card payments of 12 cents. Like that seems very inefficient. And so I think that was partly what prompted the need for the efficiency that that intellectual brings to the table. So. I think Codius has always been this sort of idea that everyone's still very interested in.

It still seems like a very good idea, but it's taken a bit of a backseat compared to IntelliJ because IntelliJ seems like the thing that you need to enable it. Currently, there is a bit of work going on on Codius here at Coil. You know, we haven't shut down entirely. We've, you know, spun down quite a lot, but we still have some, some things going on. And one of those things is. So I don't want to say too much yet because it's obviously still very early days.

And so, um, like one of the toy experiments that that I had done back in the really back in the day was like integrating it with torrenting with the idea that torrents are an efficient way of downloading media files and other and other types of files. But the two problems with it is you have both an unfairness about like the actual torrent network, but then also the content creators don't get paid, which is why it's historically been associated with piracy.

And with something like interledger, you could actually have the original content creators get paid along with the people that are, that are helping torrent. And so that was something that we had played with as well, which relies on that neutrality and the currency agnosticism. That was another example of something that we played with.

And, you know, when, when Evan and I, when we were growing up, like, you know, there was the web, you know, and there were lots of different websites and obviously that still exists. It's just really outshined at the moment by these very large hubs that sort of have taken over the mass of, of internet use. And, um, it's kind of interesting that that is happening at the same time as some of these decentralization technologies, you know, making progress.

And there's so many projects like, uh, Blue Sky and Nostr and, and Mastodon, all these different things that are trying to distribute these. And I really wonder, like, what is the missing piece that really makes it flip over and, and makes those services more attractive than the very centralized ones. I think one of the things that you can certainly speculate might be part of that is. The ability to create business models.

Cause like, if I can, you know, be on one social media network where I get paid ad revenue and I can make money, but then I'm on a decentralized version of that, but I don't get any ad revenue. I don't get any money. Then obviously I'm going to go with the, the mainstream thing that pays me. Right. And so I always wanted there to be like an alternative where you could make a living by creating content, by publishing content.

But still not be below one of these, you know, centralized platforms or be beholden to a centralized platforms. And I think that that's still very much like people are looking for that solution. It's just, nobody has quite got the recipe, right? I think for us, like the, the part that we're focused on as the IntelliJ community is like, let's make sure IntelliJ is the best protocol for that.

And then when somebody wants to get the recipe, right, they, they reach for IntelliJ to solve that problem.

Um, and the the idea of making it such that there's like one universal payment Network or internetwork such that anyone can pay anyone for anything basically it Would make it possible to include that in a lot more places So stefano mentioned, uh the web and obviously like way more of society and life is being hooked up to the internet and The payment experience is kind of clunky still and so you that does push towards centralization with Services that can hold on to your credit card

details and then charge you little amounts every once in a while So with interledger, you could have payment experiences that are much more Seamless that are hooked up to a lot more places so that that could be ranging from like paying for transit in different countries to Uh, consuming content or any, any type of business transaction that you do instead of this discussion around like, Oh, what payment network should we use for this? You just say, here's the money and then it's done.

That's a little bit of a vague answer, but it, I do think it's a little bit hard to predict how something like that would work. Roll out or like what kind of applications people would end up with? Because I think a lot of the most exciting applications of the internet were things that probably would have been very difficult to foresee before it started

Cause ads seem kind of annoying and they seem kind of a distraction, but the issue is that what ads really do is they sort of solve in a very elegant or seamless way, the problem that like. You're getting this like small amount of value from consuming content. And there's also a lot of value in companies trying to get their product out there, right? Like you'd say you're an inventor, you know, Evan says he's an inventor and he's invented lots of things.

How does he get people to use his inventions? Right? Like he's got to market it somewhere. He's got to advertise it somewhere. Right. And so that is a very important function to fulfill. And so you can't just get rid of ads.

They create a lot of value on both sides, not just in the things that they fund, but also in the things that they market, you know, And so I think that if I had to make a prediction on how something like IntelliJ could help in the future is maybe creating a bit more flexibility in terms of maybe I want to consume ads from this source, and then I want to take the money that I made from consuming those ads and then use that to pay for the content that I'm consuming from that source.

And it doesn't have to be. All linked together, because I think that's where a lot of the friction comes from. You have advertisers that don't want to advertise on YouTube because their beautiful new product is going to be next to a not so nice video. And so then YouTube cracks down on what you can say in your videos. And so then everyone is upset because, you know, their particular community, that type of discourse is okay, but it might not be brand safe.

And so then they complain that they're being censored. And so there's like all these different things that create friction when you put the content and the, the ads very tightly coupled together. Not to mention the centralization that you get where it's like, if you're a content creator and you get kicked off of YouTube, it's not like there's an equally large long form video sharing website that you can go to instead.

It's sort of like, you know, you're going to a much smaller platform from there, you know, and so I think that the biggest opportunity is to make it so that there can be lots of little hubs and they're all kind of. Working together in the sense that like, if I, if there's one hub over here that has very relevant ads, and I really like watching their ads, I can somehow hook that into my client and then go consume the content from over here.

And everyone's kind of happy that that's, that's sort of the vision. I think there, I glossed over a lot of the complexity. So there's still a lot of work to be done before we can see that, uh, realized.

And like one of those general principles for me is that when you have a system that, you know, is very important, it's very central, such as the communication system in the case of the internet and the payments, the world of payments and economic exchange in the case of Intellidger, when you have a system like that.

And there are inherent transaction costs, there's sort of a floor, like, you know, anytime you do anything, you pay this much, you know, then whenever you lower that floor, it's not just lowering the cost of the transactions that are currently happening or the things that are currently happening, so those get cheaper, sure, but then what happens is that there is You know, a hundred times, a thousand times more transactions that weren't happening because it's just

wasn't worth it for those use cases. And so with the internet, it's like the amount of data being transmitted has gone up millions of times. You know, I don't know how many orders of magnitude, but many orders of magnitude. And then in the case of, um, trading, for example, when, when we went from, you know, people calling their broker on the phone to, you know, automated algorithmic trading, the amount of trades went up by orders of magnitude.

And so you can sort of expect that something similar might happen with, with something like Intellidger. And then once you have that, then you kind of work backwards and you say like, okay, well, you know, clearly it's not going to be a person swiping a credit card for each one of those 10, 000 more transactions, because then you would be spending your entire day just swiping your card over and over. So that's clearly not what it would be, right? So it'd have to be something more automated.

So maybe you have like different systems and protocols where like your devices are making payments to each other. You know, like you're traveling and you need internet access and you don't have any relationship with a local provider, but your phone has some money. And so it just goes out and buys a internet connection from somebody.

And maybe not even a long running subscription, just like as long as you need it while you need it, it just pays a little bit, things of that nature, um, might be unlocked. I think one. Thing that while being a buzzword right now, I think is very relevant here is, is sort of AI and computers having more dynamic things that they can do, right?

Like previously you had some, it's very rigid algorithms where like some programmer would go and build things in a certain way, and then it would, if you could write executed like that. But that is really opening up with a lot more possibility today. And so I could see a lot of payments being made by some kind of, you know, AI type actors. Like, I don't really like the term AI cause it's for reasons that like, it's not really artificial and it's really just a different kind of intelligence.

But I like the idea that like, we would have this vibrant marketplace of, of computers paying each other. That's, that's, um, something I think would happen. There was actually a tweet recently by, I think it was Andrej Karpathy, who's a well known person in the AI space, that a lot of people sent me because it was something to the effect of like, if you could lower the transaction costs, then the volume would explode, which I, he's pointing to that exact same principle.

And this is the amount of units per currency that, that are expressed. And this is your address and a bunch of other things. If it could just detect all of that automatically. And so that's what ILDCP is meant for right now. It really does only information about the currency and your address. But there is a lot of other stuff that it could communicate. One of the things I ran into recently was the maximum size that an intellectual packet could be.

That could be something that I'll, DCP tells you. It could be, like, what is the nature of the relationship between you and the person you're connected to? Are they sort of your provider? Are they an equal peer? Are they a client? Um, so you can sort of figure that out. So there's a lot of stuff like that, that I think could make intelliJ a lot more frictionless to use and set up and kind of, you know, make everything kind of feel a bit more automatic and magical.

If we implement it to all of that.