Epicenter - Learn about Crypto, Blockchain, Ethereum, Bitcoin and Distributed Technologies - Ted Blackman & Gary Lieberman: Urbit - Decentralized Computing Platforms

Episode Date: December 21, 2022

When discussing the widespread adoption of true, peer-to-peer decentralization, one could easily identify various social and political obstacles. However, there is also an often overlooked technological impairment as the underlying architecture of computer networks was not designed to achieve global-scale decentralization. Ideally a complete restructuring would be necessary, from applications and programming language, to as profound as rewriting operating system kernels. This is what Urbit is trying to achieve; a decentralized network where users have full control over their computer, running a unified operating system, where applications function seamlessly, without centralized chokepoints.We were joined by Ted Blackman, CTO of Urbit Foundation, and Gary Lieberman, the Chorus One Team Lead for Urbit, to discuss how Urbit reshapes the way we interact with computers, its current development stage, and what challenges it has to overcome.Topics covered in this episode:- Ted's background- What Urbit is building- The disruptive potential of Urbit- Gary's motivation behind joining Urbit- Urbit's impact on user experience- Current development stage and potential limitations- General timeframe for Urbit adoption- Other use cases for Urbit- Implications Urbit might have for crypto (synergies)- Privacy vs. AccountabilityEpisode links:- Urbit: https://bit.ly/3A9COrX- Urbit Foundation: https://bit.ly/3G8KZXR- Ted Blackman: https://bit.ly/3Wuqyde- Gary Lieberman: https://bit.ly/3YJcRt6Sponsors:- Tally Ho: Tally Ho is a new wallet for Web3 and DeFi that sees the wallet as a public good. Think of it like a community-owned alternative to MetaMask. - https://epicenter.rocks/tallycashThis episode is hosted by Brian Fabian Crain. Show notes and listening options: https://epicenter.tv/475

This is Epicenter, Episode 475, with guests Gary Lieberman and Ted Blackman. Welcome to Episand, the show which talks about the technologies, projects, and people driving decentralization and the blockchain revolution. I'm Brian Crane. And today I'm speaking with Ted Blackman. He's the CTO of the Erbit Foundation. And also with Gary Lieberman, he works at Course 1 and he leads our Erbite team. So once again, we're going to speak about Erbit. but just be briefly before we get into the podcast, let me tell you about our sponsor this week. Our sponsor is Tally Ho. Tallyho is an open source wallet, redefining the wallet as a public good. With Tally Ho, you can safely connect to DFI, Web 3, plus a lot more.

You can view your NFTs in the wallet, across Ethereum, Polygon, Optimism, Arbitrum. And they have ledger support, so you can swap between assets and view all your account balances in the portfolio tab. They're also running a layer two adventure that rewards users for exploring the Arbitrum ecosystem with Tallyho. You can get a space dog NFT and be entered giveaways for other NFTs. So head over to tally.cash to check it out. All right. So thanks so much, guys, for coming on. It's great to do this episode, really looking forward to it.

Thanks for having me. Yeah, let's start with you, Ted. Tell us a little bit about yourself. and sort of like how you found your way to where you are at this point. Sure. Yeah. Well, yeah. So as you mentioned, I'm the CTO of the Erbit Foundation. I took over in this role about two months ago. And for the five years before that, I was working at Talon as a core dev on Erbit. And so the first four years just writing code. And then the last year, year and a half or so, I was also managing a number of people working on the kernel.

And, yeah, so I started working on Erbit full time in 2017. I had first heard of the project in 2014, but didn't understand it. And then in 2016, the docs were a lot better, and I started looking into them. And by that point, I had had enough varied experience, writing code at a lot of different startups. as a founder and as a several and as an employee at many others. And I've been doing that for 10 years, starting in 2006, my first term in college. And I sort of lucked out that I actually graduated considering that. But yeah, so I've been working in startups for a long time.

seen a lot of different things, robotics, distributed systems, web programming. And so I had enough breadth of background to understand why Erbit is interesting technically. Not very much depth of background in that much, actually. So a lot of learning about networking and operating systems and programming languages. I've had to learn a lot of that on the job working on Erbit. But that was what interested me in the project primarily was just I encountered it. I looked at it and thought, this gets a lot of stuff right that I've never seen gotten right before in program.

In a way that just the basic ideas, like what is that you build a program out of, how those programs are situated, how they communicate with each other, what they do, how the networking works, how identity works, how the operating system. all this is so much cleaner than anything else that I'd ever seen. And I still feel that way. And that's why I'm still excited to work on Erbit. So yeah, so that's kind of my, you know, the recent part of mine story. And the, and now at Urban Foundation, so we've just switched ears to doing a lot of

core development in-house. And so we're hiring a number of guys to, like, we have hired a number of people. and we will be hiring some more to work on, to do more core dev. And so expanding the size of the core dev team, that's to work on the Erbit OS itself, all the different parts of it, and really to push it over the hump to become a bulletproof consumer product. Cool. Thanks so much for that. I want to start, I want to also kind of get into something else briefly.

So, you know, most of the listeners of this podcast, you know generally talk about crypto we have done some orbit podcasts before we did one in 2017 with galen uh which i really listened to it not long ago and it's still pretty current so we'll link to that in the show notes if people want to check it out and then i think we did another one this year with josh leeman of the urbid foundation he's the executive director of the irid foundation so there's like a little bit but i still i think most people right erbit is not easy to wrap your head around and I think most people kind of still struggle of that. So it would be great if you could sort of describe for, let's say, for this kind of audience,

you know, that kind of gets crypto, that gets things around that, but maybe don't know about Erbit. Like, what is Erbit? Sure. Yeah. Well, I think for a crypto audience, I would say, you know, one of the foundational, part of the foundational ethos of crypto in general is not your keys, not your coins. Right. So you, you own your assets fundamentally through control of a private key.

And one way to think about Erbit is that we're extending that not just to money, but to all of computing. So you own your computer with a private key. You own that computer's identity on the network with a private key. And then you have full control over that computer. So what data it stores, what programs it runs, how it communicates, with whom it communicates. And then all those apps that you install into it, the programs that you actually run on there, are designed to run in a decentralized manner. So there's no central server, no central point of failure, no central choke point. And that this is, so it's sort of one way to think.

think about Erbit is it's trying to build the same world that the rest of crypto is trying to build, and it's building all the pieces of that world that aren't on chain. Because you don't actually want to stick everything, all logic and all data, onto a blockchain. It's not the right solution for everything. It's the right solution for anything where you need Byzantine fault tolerance, right? So you need a global consensus on who owns what, or even global consensus on which, you know, which transactions have been performed, right? For, you know, if you're sharing, if you're all sharing the same computer, like in Ethereum, right, you're all sharing the same computer. You just need a way to guarantee that we all share that same computer state, even if, even if it's adversarial.

But that's only a subset of the world of computing. And for everything else, where you don't want or need that kind of Byzantine fault tolerance, that's where Erbit comes in. So that's for like, I want to store personal data. I want to use chat. I want to do file sharing. I want to do video streaming. For any of those things, yeah, you want the world that crypto is promising. But blockchains can't do it.

Individual applications can't do it very well because the current stack wasn't designed for a decentralized world. The modern internet, Unix, all these things, they were not designed to be a sort of multipolar decentralized. system. And so the urban thesis, the core of the urban thesis is that, which is actually older than blockchains, but that that thesis is that fundamentally getting to a decentralized world, that same world that crypto promises is a technical problem. And it's a technical problem that needs to be solved by an operating system that is designed from the ground up to support peer-to-peer applications. Thanks. I really appreciate how you explained that. I think that's a very nice way of phrasing it.

And I do think that hopefully works pretty well for people coming from a crypto. Cool. I hope so. I've been trying for a while. You know, Gary actually has his own way of describing it that why don't want me to put you on the spot, but I've heard you describe it very well. I actually don't know if it, I mean, it could be explained better probably. I don't know if it really needs to be elaborated necessarily. I mean, I think Erbit kind of is designed as a unified system. So in the context of like if you're trying to organize your digital life, a lot of people do this by like only buying Apple products, for instance, or only using, you know, Google App Suite for all of their productivity. And, you know, there's all these ways that you can try to unify the system that you use

so that everything you use kind of makes sense with each other. And I think part of the Erbit's thesis is that things actually make much more sense if you have a system that works for everybody, for all purposes, that is truly a universal language that people can come to consensus on without any compromises. Yeah. So that's kind of like the idea of this universal computing environment and I guess what ties in there is that a whole bunch of stuff that today I guess don't happen at the operating system level do happen at the operating system level in orbit, right? So like, let's say something like authentication, whereas, you know, today, okay, you want to sign in to some application, then you can use like Google, sign in with Google, or you can

use signing with Apple. But those are like separate kind of ecosystems and they're not quite compatible. Whereas in orbit, you have like orbit ID and then that is a, you. universal thing that anyone can access. Yeah, that's right. And I mean, I think it's actually fairly obvious in the case of like in a hypothetical world where there was only Erbit and, you know, Erbit had no incumbent that it was trying to argue against. It would be kind of obvious, I think, that identity is something that you only have one of. It's kind of baked into the word identity that you have one. And so a lot of things that don't work properly in like the Web 2 ecosystem,

And even Web 3, actually, the reason why they're not quite compatible is because the identities just aren't compatible with each other. And so if you have a unified identity system at a very early point in the stack, a lot of problems just go away. They kind of, it's not even that they're solved. It's that the preconditions for the problems never even come about. Yeah, that's right. Like most people don't know, for example, that IP and TCP, you know, the protocols that the Internet uses, don't have any concept of authentication or encryption. Those have to be bolted on at higher layers.

And because of that, you get all kinds of issues. You get a whole field of IP security, IPSEC, DNS security, DNSSEC, right? Like, these things basically shouldn't exist. They should be in the network protocols. But these network protocols are 50 years old before anybody even thought to use encryption. And before there were, you know, when there were maybe 10 people on the internet instead of billions. So yeah, so with Erbit, those things are built in at every pack is authenticated and encrypted. One question actually when Gary and I were talking a little bit about, you know, the outlines.

And there's this question that I think is a good question to ask right here. So, you know, Erbit describes itself as this clean slate. And, you know, you talked about these like global protocols like TCP and IP. So can you talk a little bit about, you know, how much of kind of the existing internet and computing stack? Like how much does Erud like throw away and to what extent are existing technology still being used in urban? Well, at a basic level, it throws them all away. And but that's not the whole story because despite that, we know that Erbit must exist in the present and be useful for people in the present.

And so Irvitt also has the capability of acting as a web server to the old web and making web requests, HTTP requests. So this allows you to access it from your browser, access it from your phone, you know, have it serve a blog for you, right? So the studio app from Terrell lets you post a, host a blog. blog using your Erbit and even they're working on adding a paywall to that too so you can even take payments through your Erbit so it Irbit is not basically Erbitt's building a whole new world unto itself right everything from machine code programming language operating system kernel set of applications network protocol identity system all of that but that new world can't communicate to the old world because that's very important in practice.

Cool. Thanks for elaborating that. Well, maybe we can go to Gary also briefly. Maybe you can tell us a little bit about you and how did you get into Urban? And what was the thing that captured your attention when it came to Urban? I don't think my background is particularly interesting, but I was before, I mean, I knew about Erbit for a couple of years before I got really into it, and I underestimated it for a very long time. I would say what got me really into it was struggling with two things that I didn't quite understand were related. One was that I was trying to organize my digital life in a way that really made sense to me and optimize my workflows, kind of just running on the assumption that I realize I'm

much happier and much more productive in my physical life when I'm well organized, so I should be able to apply that same logic digitally, and just coming up against multiple blockers everywhere, where it's like, okay, to actually solve this problem, it's like a full-time job for 10 people creating and maintaining a system that actually does what I want it to do. So that was one thing, and also just noticing on the other hand, I mean, it was right after COVID, just how dystopian Web2 was. It wasn't surprising to me at all that we had like this huge increase. and like political violence and transgression

at the same time that everybody was stuck in their houses with nothing but web two to entertain them. Like it seemed very clear that that was like an insanity-inducing situation. And Erbit, I think successfully made the case that that was also downstream from the same technological problems that led to my own inability to organize my digital life in a way that was automated and made sense. And so I gave Erbit the benefit of the doubt enough

that maybe if I learned something about it, and I spent quite a lot of time learning it, that I could kind of test the assumptions and at least make a case for why it was or was not a waste of time. And by the time I was done and felt satisfied that I understood it, I was totally unable to recreate the conditions that I had before where I was not a complete urban maximalist. I can't even remember what it's like

to believe that there's any other possibility for the future computer. Fairly common story, actually. You know, there's this question. I kind of wanted to get to at the end, but I think we should get to it now. So, you know, we've talked about, you know, we've touched on urban a bit, right? Like what's different?

And, you know, you guys have both alluded a little bit as to why you're excited about it. But if we, if we now kind of think forward, And imagine that Erbate, you know, really accomplishes the things it's trying to accomplish, right? Like, so it really, I get, so it, you know, fully realizes its ultimate potential. What does that look like? How long does it take and what does the world look like? What's different for people in the way they use technology, in the way they live their lives,

in the way political systems work, you know, like, what is, what is, what, the kind of visions and ideas that you guys have around this? I can start with a few things. So one is that I think you'll end up with a lot more, well, localism is one word. And you end up with a lot more subcultures that are sort of, each of those subcultures will be autonomous in a way that they aren't really digitally autonomous at the moment. Or if you're running a subreddit, the subreddit might just get shut down. The mods could have a bad day And then you're not there anything

And that culture can be gone You can be a Twitter poster And then You know, big following And then put your foot in your mouth one day And your account gets flagged by an AI And banned And then years of work that you put into developing that audience

And you know, sort of building a whole network Of personal connections based on that can just be blown away in an instant by essentially a mistake. And so this actually kind of puts a damper on the ability for subcultures to form. And so you end up with more of a monoculture, right? So I think Trent from Holium said Facebook is 3 billion users in no culture. And that really stuck with me. So I think it's not entirely true.

There's some, right? But I think it's what we'll see is a lot more of those, a lot more subcultures that are able to flower more, right? They're able to develop their own art, their own ways of communicating. They may eventually physically co-locate. They may do a biology-style network state. They may do something else. I think that's sort of digital, like the ability for people to coordinate, collaborate online has been hand. by a lack of ability to, for people to control their own software.

Right. So, you know, you're probably in some signal group chats. And not just signal, scattered all over the place, really, right? But for any one of those, you can't, you can't say, oh, well, let's, let's fork the signal client. Right. Like, let's have it, you know, integrate with something. Right. integrate with a calendar even or something basic, right?

There's just no way to do that because that's its own world, right? It's siloed, right? And so basically that sort of limitation where you kind of can't just arbitrarily extend things. You can't arbitrarily get your programs to work with each other. That's a big cause of what Gary was saying that he ran into. But it also just prevents like groups, social groups and other kinds of organizations from being able to really make the most out of networked computers. So Herba sort of flips the whole situation on its head

because right now you have to go through some big central server to do almost anything on the internet. And instead, it says, no, you can download an app into your Erba, run it yourself. And you have full control over that yourself. And then you and people you want to collaborate, with have control over that that small network. So it's much more bottom up.

And so I think like we're really not used to seeing bottom up organization digitally. But I think that what it's going to look like mostly is, yeah, better art, more art, better, like I think more seamless economic activity, right? Like a lot more people starting small companies, starting small, you know, little investment co-ops, starting local businesses, things like that. I think that, and getting local investment, right? So all sorts of things that, and all kinds of different modes of interactivity that are very difficult to build right now, right?

So, you know, if there's an interesting example of this, which is an app called Radio on Urban that was built a couple months ago. And it's sort of like a jukebox for video. So you all get into this room and you can set which video is playing. and then in queue another video and then people can talk about it as it's going, right? And, you know, it's not an earth-shattering application, but it's something new.

That's something that you can't build otherwise. And so it's just sort of one little hint as to what sorts of, like you can just have this, you ended up lots of different kinds of interactivity, different kinds of bottom-up social organization that are infeasible in the modern computing setup. Yeah, and just one thing to add to that, I think, I have key importance is the idea that within a unified system,

a small-time app developer can actually have a big impact because they are working within the same language as the rest of the world, meaning two things. In terms of culture, there can be an application made between friends and a group and not have this, you know, this big hump to overcome to actually get it usable between everybody. If I develop an app for just the people in my friend group, it would be very easy in an urban world to just have them download it and try it and use it. And it wouldn't necessarily have compatibility issues with the rest of their digital life. And that also makes a lot of sense in a business context.

So if you're running a business and you realize that you have a very lucrative small business model, and you're asking the question of, you know, what would it take to franchise this business? Probably the answer to your question includes, you know, you need some kind of software system that, you know, that captures the entire business logic of your operation. And this is in practice not possible for most small businesses. You have to overcome a significant financial barrier to do that kind of thing. But on a unified, you know, a software ecosystem, it actually becomes quite a reasonable proposition. I wouldn't call it easy, but it's much more accessible to simply capture the logic of a business within one program,

and that I think would lead to exponential productivity gains in pretty much any industry. Yeah, just to elaborate on that a little bit, so there's one of the big differences between writing code and Erbit and writing code for the normal Internet is the digital. is the distance between writing a toy and writing a production application. So normally, in the normal internet, normal computing, a programmer can take a weekend and write a toy. And that toy would be usable. It has the business logic, right?

The sort of meat of the application is there. But it's nowhere near something that people could use. So you can't just sort of have your friends use it, have it work, have it be something that is actually part of your lives. In order to get to that stage, you have to do a lot of DevOps. You have to have some sort of authentication system, maybe a few with backups. What happens if somebody forgets their email address or whatever, right? I mean, you have to administer a bunch of servers.

So you have to have some sort of cloud infrastructure. You may find yourself researching Kubernetes and Docker. And there's this whole long list of tools that, that the modern programmer has to be familiar with in order to actually deploy an application to users, even if it's just a web page. And so it's that difference between the meat of the system and all the sort of rigmarole of deployment. That's what brings in a ton of extra cost that makes it prohibitive to just say, hey, you know, we've got a few friends here. Let's write a little app that does this thing that we needed to do. And so this is, this is,

This is true already, right? So a good example of this is Justin Murphy commissioned somebody to write an app called page, and it lets you copy some HTML into your orbit and serve it as a webpage. Because he just needed to serve some simple stuff. So he asked somebody to build this for him. I think he paid him $500, it took a weekend. And now that app exists, I installed it on my orbit.

It works fine. right so I was able to download this run it myself on my own herbiv right from $500 worth of work and that's that's a fairly unusual setup and we're going to make that even better in the future so we'll reduce that cost even further

and that difference will continue to increase so that's I think one of the things that's most exciting from a developer standpoint it's like yeah you can about urban is that you can write something publish it literally in a weekend and it's real. It's already real. Yeah, I mean, this is one of the ways, if I explain Ermit,

that I also tend to emphasize is, you know, if you're developing a web application or some kind of application in normal web, that because you have these additional things that, okay, now I'm going to have to run some infrastructure for it and the cloud and, well, what does it mean? I mean, I have to make like an AWS account and now I have to pay every month. And if there's more users, I have to pay more. So now all of a sudden I have to run a business, right?

And I have to have revenues to do that. And maybe that means I have to raise investment or I have to charge people. And maybe I didn't want to do that. I just wanted to create something, right, that like people wanted to use. And in urban you can kind of do that, right? they could develop some simple application. Maybe people love it and a million people are going to run it on their, you know, on their orbits.

And I don't have to do anything, right? I don't have to charge anything. I don't have to pay for any infrastructure. And I think that's just going to be so attractive for developers. And it's going to like unleash such a new class of applications people will build. people play around with and so much creativity that becomes possible when you when not everything has to be connected with some business that you have to build yeah i would say generally there's a very large class of questions that begin with why can't i just that erbit um allows you to just

and and shipping code to somebody's computer is is definitely one of them why can't i just build an application you download it for me and run it another one is why can't i just why can't i just send you a file. Why do I need Dropbox? These are these are just questions that you only have to ask because of the kind of insanity in computing and urban I think is kind of fundamentally trying to provide that sanity. So let's talk a little bit about, you know, what is Erbit today? You know, what's the current state of the technology and what are the limitations of the system? Sure, yeah. So today, first of all, Erbit does run. Sometimes we get asked that still, but it's been running since 2013. And the, so you can get an Erbit. You can buy one off Ethereum. And so you can buy an Erbit ID. Once you have that Erbit ID, you can boot ID, you can boot your ship, which is the name for your node, so you can boot your orbit. You can run that. it yourself on your laptop or pay one of a few different hosting providers to run it for you.

Or you can run it yourself in the cloud if you're willing to do a little bit of system administration. You can also move it from any of those three places to any of the other ones. And Gary actually has been working on standardizing the protocols for how that can be moved from place to place. But you can do it yourself already. So, Erbit, you can get one. You can run it. And you can install apps into it. And it comes with a suite of apps. The flagship apps are groups and chat. So you can chat in groups.

It looks a lot like Discord. Or Slack, something like that, right? It's a chat app. It's sort of asymptotically approaching feature parity with those. There was a recent update that made it a lot closer, actually. Now it has threaded conversations. and emoji reactions. It's starting to feel very similar to any other chat app.

So that's mostly what people use Erbit for right now. But you can also install any number of other apps, and there are about over 80 of them that have been published publicly. It's hard to know how many of them exist total, but there are many dozens that you can try out. And some of those are more important than others, I would say. So for example, there's one called Campfire that lets you do video chat.

That gets over WebRTC, but it's negotiated by your ErBitt, so it's fully decentralized peer-to-peer video column. So, and there are all kinds of other things that people are building. So that's where Erbitt's at. And the chat stuff works pretty well. There are some rough edges.

I would say that the big limitations have to do with scaling, and hardening, and some with developer experience. So as a user, basically, you know, if you join a group that has more than a few thousand members, and there are a few of these groups, then it might be kind of slow. It works, but it's a bit slow. And if you're publishing a group that has many thousands of members, then you generally have to do some maintenance work on your server to make sure it doesn't run out of memory. So, you know, Erbit is supposed to not require any maintenance. from the user, it should maintain itself.

So it's not quite there yet. And in particular, with scale, right? So if you're supporting a lot of subscribers, then the system doesn't scale as well as it should. And so you have to do some maintenance work, and things get slow. And so that has to do. The reasons for that are basically twofold.

Well, three. There are three reasons, essentially. One is that there's a limitation to how much data you can store into orbit right now. There's also, it just doubled, but it's still limited. So there's a limit to that. There's a limit to the amount of network bandwidth that your Erbit can take advantage of. So we're working on that as well.

And then finally, Erbit's programming language, knock, the machine code is slow. So the combination of those three things, makes, put some scaling limitations on the system. All three of those are under active development, and I don't expect any of them to be show stoppers, but, but that's, that's the status. So the other aspect is hardening. So basically there's this refers mostly to security, some to reliability. So there's some reliability issues.

There's still some bugs. Sometimes I think crashes much, much less frequent than it used to be, but it still happens. And so getting the system to be really bulletproof, reliable, that's one of the big things we've got to do. And a very closely related phenomenon is making sure that it can't be hacked into. So making sure that it's secure so that people can't get into your orbit and download all your chats. We want people to be able to store private keys for relatively small amounts of money. But you should be able to store private keys in Erbit and insert anything other than like your life savings. It should be relatively safe.

You should feel secure about it. We're definitely not there yet. So as a caveat, like if you go play with Erbit, keep in mind, it has not been audited for security. There's a good chance that there are vulnerabilities in there. So treated accordingly. But that's also under active development. So there's a hardening process that we're taking the system through. And the other piece of that, I would say, is.

protection against denial of service attacks, which is, I think, going to be a sort of long tail of hardening to make it sort of able to resist larger and larger and more and more well-resourced groups of people who are attempting to just shut down the network, or at least shut down your ship on the network, your node. So there's a lot of work to be done on that front as well. Also not, I'm not particularly, you know, worried about it, but it's a fair amount of work in front of us to do that. The third remaining sort of limitation, I would say, is in developer experience. So basically there's when it's a pretty new world, right, you've got a whole new, whole new machine code, whole new language, one operating system, whole sort of paradigm of programming. And all these things are, you know, 10 times newer than the existing world. And so there's a lot of tooling that Erbit doesn't have that other systems have. So a lot of the, so that, you know, this developer experience is very good.

already in some ways, but in other ways it needs a lot of work. And so a lot of that has to do with building tooling, building debuggers and having a better, you know, printer, pretty printer for data and a lot of other technical stuff. But basically making it just a lot, building a lot of tools for developers. And there's a deeper piece of that, which is making sure that the APIs that are presented to a user space program or somebody writing an app to make sure that the way that you interface with the operating system is clean and easy to use. And in particular, easy to, it should make it easy to write a program that's correct. And so there's a big problem with this right now, which has to do with subscriptions. And so there's a whole sort of complex of projects that

we'll be working on over the next year to called subscription reform to redo Erbitt's subscription system. So how you synchronize data from Erbit to Erbit, that should get a lot simpler, cleaner, and more scalable as well. So, There's roadmap.urbit.org has, that's a pretty new thing where we laid out the technical roadmap for Urban. So if you're interested in going deeper on the technical details of exactly where the system's at and what needs to be done to improve it, that has some pretty detailed explanations and linked to ongoing work on GitHub. Cool, thanks. Yeah, I mean, with Gary, we were actually looking a little bit at the roadmap the other day. And I have to say, it's like very nicely presented.

And so we'll also link to that. So if people want to go deep and check it out, what are the timelines on this? Like when, when is orbit? I don't know. Do you see, do you see kind of different stages of readiness of orbit coming up for different types of, I don't know, degrees of adoption? And what does that look like? and yeah on what time frames I think this is a little bit of an aggressive

time frame but I generally say around two years to get to having a to have it be bulletproof and scalable enough and reliable enough basically so there's the system will continue to improve basically forever

the kernel itself will actually stop improving at some point the version numbers go down to zero. We call this Kelvin versioning. And this is to standardize the system. So all the different parts of the kernel each have a Kelvin version. Those all go down to zero. Knock the machine code is at Kelvin 4. And Gary and I actually did a podcast where we talked about what might cause it to go to version 3. But at the moment we don't actually see that happening. We think that's probably done. But the kernel, different parts of the kernel are at versions 140 for the language, up to about

400 for the standard library. So there are a finite number of versions that we have to play with there, which will take somewhere between 20 and 200 years to really congeal all the way down. But that's the kernel. There's also the runtime, which is the program that runs your Erbit, and that is not Kelvin version. And so that will continue to improve forever, but that's basically just making it faster and faster over time. So there's no sort of fixed point, there's no like one point where we'll say urban is done exactly um but uh there's a point where we will be able to say this handles all the kinds of applications that we want it to handle uh and i think we'll be able to handle most of those by two years from now and so that includes having being able to

confidently say this thing is secure nobody's going to be able to steal your data out of it exfiltrate your data um it'll scale reasonably well so if you want to have a million people in a chat channel. Two years from now, you should be able to have a million people in a chat channel. So basically, it should approach the level of scaling of Web2. And that's an interesting set of problems. Most people don't think of decentralized systems as being capable of scaling up to the size of Web2, where you need many data centers to run it. Erbitt actually is designed to do that. So there's a very interesting set of discussions to be had about how it plans on doing that. But I do think it's

possible, and we're on our way. So it'll scale some. It'll be quite, it'll be quite secure. It should be quite reliable. The program should not crash. And it should be relatively easy to write. It should at least be, yeah, it should at least be relatively straightforward to write an application that synchronizes data correctly without having to worry about correctness. So like where it gets where it's a correct application without having to think too hard about it. Now, by two years from now, it may still have quite a bit of boilerplate, right? It might be kind of awkward, clunky to write those programs. But when you write them, you should be able to very quickly read them and say, yeah, okay, that's definitely right.

We're not there yet, and mostly what's missing there is the interface to subscriptions, the interface to synchronizing data. There are a few other things that aren't great either. So I don't expect that that API, right, like what it takes to write an application will be really, you know, smooth and just, you know, really polished by then, I think that'll take another couple of years to really get, to really get nice. But it should be possible to hook in to scalable content distribution from an application so that you can support, you know, many people subscribing to you, like in the millions. And to feel confident when you write it that it's correct and that the kernel is

handling all of the sort of hard synchronization problems and other correctness issues for you. I was kind of wondering if there are any applications that you would like to see on Erbit that it's just not ready for because of, you know, the maturity of core development and, you know, when do you think it'll be possible and what it will look like? Tons. Okay. I want to store all my files in there. I have hundreds of gigabytes of music recordings and photos, all kinds, you know, hundreds of papers that have collected over the years that I want to read, computer science papers, ebooks, tons of stuff. And I actually, I had them in Dropbox for a long time. Then I decided to stop paying for Dropbox. They pulled them onto my computers.

And I have like 300 gigabytes of just crap sitting there on my computer. And it would be meaning to to upload it somewhere else and I've gotten around to it. And that's, that's a lot. should all be in my orbit. And not only should it be in my orbit, but yeah, I should be able to share that with you very easily. So in order to share media with you through orbit, you need much higher, so first of all, I need to be able to store that data in urban. It's too big right now. I need to be able to stream that to you or you need to be able to download it quickly, so we have to be able to max out a network connection, right, for the bandwidth. So we need a better network protocol. And I need to be able to publish that to a lot of people, so we need scalability of the network protocol.

So there's that, right, anything with file sharing, podcasting, you know, live streaming. You should be able to do all that from your orbit, right? And like, imagine doing a live stream from your orbit. It's fun to think of all the other stuff you could build in there, all different kinds of interactivity and, you know, integrations with other other urban apps. And then, yeah, the other thing is I want to be able to store a lot of my private keys. in urban. So I think like, yeah, for for stuff that's like where you really want to keep it cold, then okay, yeah, keep it cold. Don't put it in your orbit. But, but I want to have sort of what, where I'm paying my bills from, right? And where I'm, you know, sharing expenses with friends.

You know, an urban version of, what's the app called where you like, you all go to a restaurant and split wise, right? My friends and I use split wise for go to a restaurant. So it lets you settle up later. I think there's another one too. I think there's multiple. I bet there is. Yeah. Yeah. But anyway, like that kind of app where it's like, yeah, here's how I handled my sort of day-to-day expenses, income, et cetera, right? Like that, I should feel comfortable storing that in my urban, right? And then not only that, but then from there, I can do things like have urban integrate with the lightning network, have it integrate with, you know, the Aztec protocol on Ethereum, have it integrate with Gnosis safe, safe, I suppose.

Right? Have it integrate with all this different stuff. So, but in order to do that, I have to be comfortable having some hot wallets inside my earth. And so for that, it needs to be secure. So there's this whole sort of big playground of all these things that could be built, right? And it's like, those will be toys until we make Earth, but secure, and then they can be done

for real. And so I'm very much looking forward to those classes of values. applications. Yeah, I'm glad you mentioned the podcasting thing because I think this is actually like a very powerful example, right? So for those who like you don't know basically if podcasts or podcasts are basically distributed via RSS feeds and so as a podcast, you know, you're putting up this podcast recording and then people download it and you don't and there's a bunch of limitations that come with that.

One is that you don't have any way of messaging those people. Like if you want to build like a community of your podcast listeners, then you'd have to make this cumbersome thing of like, oh, you're listening to this audio. Now go and click in the link in the notes to go somewhere else. And then we do this community. So but most 98% of people, they're just like listen to it. And because it's so cumbersome, people don't even try.

And of course, charging for podcasts. is also basically not possible, or it's only possible through these, again, very, very terrible hacks that people do. So I think this is something where, you know, if you can just distribute your content, message everyone who's listening to it, maybe have better data about it too, because I think this is another thing that podcasting is very limited with. Although I guess in orbit, maybe this is also something orbit will be limited with, no, since people... What do you mean better data? Well, I mean, in podcasts, like, you don't know, for example, do people actually

listen to it or not? You only know they down the analytics, tracking, that kind of stuff. Yeah. Well, I think people could hide the fact that they've listened to your podcast relatively easily, but most of your listeners probably wouldn't. So if you kind of requested that information, I think most of your listeners would give it to you and you'd have a reasonable idea of what your audience looks like. Yeah, you can also track the networking requests coming into your orbit, right?

So you can track just how many requests came in. Or I should say, you could build this sort of analytics to track that sort of thing on your orbit. There's a little bit of that in the urban runtime right now, but it's not as granular as it could be. But yeah, you could build a system like that so that anytime somebody tries to fetch this, well, they're fetching it from my orbit, right, so I can tell. You may not know that exactly if we end up building a sort of,

distributed caching system, which is likely. But yeah, so yeah, Erbit probably requires a little bit more work to get really good analytics, which has to do with sovereignty, actually, right? Like there are some places where you have fundamental tradeoffs between, you know, privacy and sovereignty and, you know, being able to track what people do, which is the opposite of that. But, but yeah, I think there's quite a bit of middle ground in there. What about you, Gary?

I mean, you mentioned the example that I really liked and I hadn't actually fall off in this way before of having this business that is in a code and you can kind of franchise it easily. So I think that was a nice example. But what else? What are some other use cases for a bit

that you're most excited about? Actually, I think what I would find most useful are things that are currently possible and just need a little more developer effort. So I'm very excited about Ngram, a whole-eum product, which should work towards

kind of replicating the feature set of notion on Erbit. I personally find that if I'm journaling, even if I pretty much trust that I'm not going to get snooped on by whatever third party owns my data, I'm just going to be a little more hesitant to really speak freely about what I'm thinking if I know that it's not private. And so I would much rather have something like that on my orbit. and I personally prefer to journal on my local computer rather than a more powerful platform like Notion for that reason. And I think Herbert can do it. I think it's just a matter of developer effort and kind of getting the UX right.

And another thing is task management. I use Task Warrior, which is a Linux program, and it works really well. But I would probably prefer for that to be on orbit as well because I would have better sharing functionality and I wouldn't have to work within two different systems. You don't use Emacs? No, I don't use Emacs. Task Warrior is good enough. I don't know.

But no, I mean, these are kinds of ways where having a personal server makes a lot of sense, because you have a variety of devices. I'm not going to carry my laptop necessarily to all the same places that I might carry a phone or a tablet, but I still want access to these basic things, like my notes and the things that I want to do. And these are low amounts of data.

It's just important to me that I can access them from anywhere. And the only way to do that in a sovereign way right now is to deal with honestly just a very terrible U.S. And, you know, do probably, you know, more Linux work than most people want to do. And I think orbit is going to make that much more accessible and, you know, more free. Well, let's talk a little bit about, I mean, we've touched on it in endpoints, but about the kind of crypto-urbit synergy. So once orbit, you know, let's say two years from now, we have addressed some of those issues around security, scalability, you know, you can keep your keys in orbit. It's pretty safe. What does this convergence of crypto and orbit look like? And, you know, what are the most

important implications of that? Yeah, I think there are a lot of implications from it. And the convergence happens in several different ways. And so there's actually a lot to talk about here, I think. So one is, let's start with some basic ones, right? So it's basically like anytime that I as a user want to interact with the chain, I should use my Erbit to do that. So even if that's something simple, like I want to use Uniswap, right? There's an app on Erbit that serves the Uniswap front end.

but I should be serving that to myself and Erbit has very good properties for code signing. So the software supply chain can be secured much more effectively on Erbit than through almost anything else because we have a real PKI or a real public key infrastructure, real identity system that actually lives on chain. Every Erbit ID is an NFD on Ethereum. And so what that implies is that if I can have a lot more protection that the front end that's being served to me is not spoofed. There's no DNS spoofing. There's no other stuff, right?

Like it's something where I can really get a lot of verification that I'm using the piece of software that I mean to use. There have been huge hacks, right, where somebody spoofed some front end and, there go hundreds of millions dollars. So, so, uh, so Irbyt, first of all, you know, that should be the place, that should be what serves a client to you, right? Some sort of, uh, any user interface you're using for interacting with change should be done through urban. Um, this also lets you do multi-party computation, right? So if you want to do, uh, a multi-sick, um, then that should be, there should be an app for that. Uh, it should be a nice app to write, you could write it right now. And, and then that, you know, then, okay, we want to,

you know, the three of us have a multi-sig, let's say, and we want to, you know, we want to move some funds. Okay. Well, we can use the Erbit app to do that. We don't have to copy and paste addresses. We don't have to worry about backups for our keys. We don't have to worry about backups for the state, right? Like if you're tracking, you know, the nonce of an Ethereum account, you can use your Erbit to store that and make sure that, and then you know, and then you know, that it's not going to get deleted, you're not going to forget it, you're not going to have, you know, non-s reuse errors that could cause security problems, right? This whole long tail, like, things that can go wrong with crypto. And a lot of the stuff that people worry about with crypto in

practice when using it's like, well, what if I copy and paste the wrong address into the wrong thing? And there goes all my life savings. It happens. Right. So what you need is some abstraction layer there's say, look, I'm sending it to you. It's just my computer, sending it to your computer. And yeah, I guess there are like public keys and addresses involved, but I don't want to think about it. I don't want to see him. Right? Like, I don't really think about that. To the extent that I have to think about that with normal banking, it's also a bit of a flaw.

Right. Like, so there are those layers. It's like, okay, I can do my own stuff. I can have my own wallet in there. I can use it for multi-party computation. And then, yeah, I should also use it for storing all the client-side state. So all off-chain state that I need. if I'm using Safe, right, to manage some assets and do other things like that.

My actual personal data has to be stored somewhere. So I should store that on my urban. Another thing you can do for you is it can act as your watchtower, right? So let's say you're doing something with lightning, right? Or that's not the only thing that needs a watchtower. But for some of these things where you're having like a state channel, you actually need some program running on your behalf monitoring the chain to see if something happens. So if something happens that's against you, this thing can do an adversarial close of the channel and make sure that your funds don't get sold.

So you can rely on somebody else to do that or you can use your orbit for that. Then there are deeper integrations. So something where basically you want to have an application that's mostly off-chain but has an on-chain component. So let's say you have a video game and there are in-game assets that you actually want in-game. me for, right? And you double spend protection on those assets. So for that, there's, you can do that just sort of normally, right, where you can write an urban app that's primarily, you know, in orbit app, but also talks to a chain. You can do that by talking to existing chains. And then there's also the Oak Bar project, which is aiming to, uh, they're building a ZK

roll up as a layer two on Ethereum. that uses NOC, Erbitt's machine code, as it's to write smart contracts. So they do ZK proofs of knock code. So this allows you to write your application in Erbitt's programming language. And you also write some little piece of it that stays on chain. And those share the same data types. They share the same helper libraries. They communicate very seamlessly with one another.

So there's also a lot less code involved in. deploying that application. So one of the big problems with deploying DAPs, as they're currently conceived of, is that you write a little bit of solidity, but most of what you write is JavaScript because, oh, you know, there's actually a lot more code that runs off-chain than runs on-chain. And so you have to still deal with all this Web 2 stuff, all the difficulties of deploying Web 2 application, along with the difficulties of deploying Web 3 application, because you can't screw up the solidity either, right? So it's actually just... pretty scary thing to try to do.

So Erbit doesn't necessarily make the smart contract part of it that much easier. But everything else around that, it's a lot easier to integrate that into an application. And then there's an even deeper layer, I would say, of integration between Erbit and blockchains. And this is more speculative, more sort of philosophical. But basically, I've been thinking about this lately that blockchain is an Oracle problem. Just in general, if any data is going to be put onto the chain as a transaction input, the data has to come from somewhere off-chain. It can't be verified just by being on chain the way that's derived data can be.

So a general, you know, so it's like a, let's say, you know, you want to pay me for doing some work. well how do you how do you know that I actually did the work and so something somebody has to put that data on chain right and that's an Oracle problem because when they put that data on chain you have to try to figure out well how does that how does how do we know that that's true that that's accurate information right so chain link has a whole system for this right I think there are a bunch of others there the chain link is for you know or Oracle for price data and so that has particular characteristics where you can, you know, you can sort of average them out and throw out the outliers and have some economic incentives to get a good answer there, right? But that only

applies for certain kinds of data. And for a lot of those sort of, a lot of what people were interested in blockchain solving back in 2017, 2018 that never, or that have not yet panned out, I should say. Things like, you know, using it for, you know, the title for your house, right, or like these sort of more real world, tokenizing physical assets. Some of that's probably just a bad idea generally, but my guess is that there are several of those that actually are good ideas,

but they're not feasible yet. And what's needed to make them feasible is includes a more generic and more, like a more general purpose, Oracle system. And so, Erbit, I will actually get to the point here. One of the ways I see here, but as a general purpose oracle system, like the lowest layer of an oracle system where basically you want to put some data on chain?

Okay, well, it comes on chain signed by the Erbit that produced it and located within Erbitt's Scrii namespace. So this is very technical concept, but basically Erbit has a way of assigning a permanent, immutable name to a piece of data and attesting to that by signing it. with using the urbid's private key, verifiable by anybody. And so it's a way of laying out all the data

that gets produced by any orbit on the network and having that erbit attest to that data. So at the moment there's no double spend protection, or there's no Byzantine fault protection on this. So there's nothing to say that I can't run a malicious urbit kernel that signs two different versions of the same data, the same path. But, and most of it doesn't,

most of the data that Erbit deals with doesn't need that protection, and so shouldn't have it because you actually, again, you only want blockchains for certain things. But you can take that data. You could write a contract that does enforce Byzantine fault tolerance on those bindings. And then basically you could use that as a way of just ingesting arbitrary data from anyone's personal computer. And it comes attested to by them, you know who they are. not, I mean, you don't know their social security number, but, but you can tie that identity to other actions that that orbit has taken. It does develop reputation. And so because it develops reputation,

that's where you can build systems of trust. You can build reputation systems, on-chain reputation systems that deal with us, that you could use programmatically to filter all this data. So this is a whole line of thinking. None of this has been built yet. It's just sort of something that's been rolling around in my head. But I think there's something fundamental there, right? Where basically blockchains are specifically for dealing with objective data. Right? Like what is objectively the sort of consensus truth about something?

Erbitts are designed for dealing with subjective data. What is my version of this? What is your version of this? And so I think these two things complement each other very well. And you need both of them to really build a sort of cryptographically sound world. You need a world that has hardness, as Josh Stark described in his essay of Adams, institutions and blockchains. As civilization progresses, the hardness of our institutions increases. Blockchain is a great job of hardening currency, contracts, potentially law.

But not anything subjective, really. And so for, so Erbit is really where you harden the sort of subjective reality. I don't know what you think about that. Yeah, I want to ask about one thing that I've been, you know, pondering a little bit. And I'm kind of like unsure how this is going to play out. Right. So today in Web 2, obviously privacy is pretty bad or maybe terrorist. terrible, right? And orbit has this, you know, fundamentally different paradigm that, you know,

it gives you more privacy in some ways, right? Because like your program runs on your orbit and your data is there. But at the same time, because everything you do on the network is like associated with your planet ID, it also seems to be the kind of this aspect where maybe there's less privacy too or like you know at least in the current system you know you can make like different usernames use different emails make different accounts and so I'm curious like how you think this is going to play out I would say there's more accountability right and that does cut both ways because yeah if you get because so basically in orbit you have this ID and it's typically a very permanent ID. People get really attached to them. I'm very attached to Ravenus Rickfer, Gary. If somebody

tiller told us away from you, I imagine you would not be very happy with this. You get attached to it the way you get attached to your sort of normal human name. It's a little odd, honestly. It's like that's part of the system works a lot better than I would expect it to. But there's quite a bit of empirical evidence for this. People get very attached to their name. So you get attached to your name because that's really who you are on the network. It's how people know. you. And I think that'll only increase as people start doing things like financial transactions and other sort of more serious, more serious uses. And so, yeah, there is a tradeoff there somewhere abstractly, right, between, you know, can you, between being accountable for your actions

on the one hand, which also means, you know, you have a reputation. So developing, the development of social capital requires attributability of actions, right? Like, you have to know, you have to that it was me in order for me to develop that reputation. But then conversely, you know, if I do something wrong, then, you know, my, then everyone will spit whenever they hear my name for generations to come and the sins of the father should be revisited upon the child or whatever, right? Interestingly, I think blockchains have a very similar problem, right, where like the provenance of transactions is actually completely visible on a chain the way it's maybe not within a

more traditional banking system. So an Erbit actually takes the same approach that blockchains do, which is you have a layer of pseudonymity. Right. So any particular Urban ID, yeah, that accrues reputation. But that idea is not tied back to your real world identity unless you want it to be. Yeah, I should mention that anonymity is possible on Erbit. People will know you're anonymous. And in most contexts, you may want to be unaccountable in a given situation, but the other people in that situation probably want,

everybody else around them to be accountable. And you would probably like everybody else to be accountable, too. So while anonymity is possible, it actually makes it more difficult to socialize in many ways. So I think just like in Web 2, on Erbit, we will see some anonymous social happenings, but the majority of interaction will be accountable with this kind of skin-in-the-game-based system. Yeah. So specifically, there's a type of urban idea that's called a comet. You don't have to buy it.

you can create an infinite number of them effectively. And yeah, you can throw them away. They're ephemeral. And so they're anonymous IDs. And interestingly, yeah, we find that most substantial groups ban them because they tend to be trolls. But there are places where they're not banned because people really want to have anonymous conversations.

So, you know, it's different use cases. Different use cases imply different things. And so you can you can do both with Erbit. So I guess there's Erbit. developers.urbit.org is the best place for engineers who want to learn more. If you want to get started using orbit as a user, there's getting started link at at herbit.org as well. So that's all very polished and user friendly nowadays. Yeah, we're still working on it too. It's getting better.

Yeah. I mean, I'm just very excited. I mean, today and also just for like, you know, what does it look like today, right? I mean, there are some of these applications, But I think right now what really works is just as groups, right? So you have some community and you can message each other in your chat and it's like pretty nice. And I think that's already something that, you know, at least for a lot of people, it's enough to kind of trigger a lot of thinking and a lot of seeing of possibilities. But I think what I'm super excited about is, you know, the point where we really have compelling applications. You know, they're like actually cool. And I think we're pretty close.

I don't think we're there yet, right? I mean, at least personally, I haven't used anything where it's like, you know, this is really cool beyond just illustrating some possibility of what could be in the future. But I, you know, I would guess that this is something that we'll get to within the next six to 12 months. And I think at that point, it's just going to be very exciting, right, especially to have people from crypto, like come over and explore orbit, then I think there's going to be, you know, an avalanche of innovation and adoption that's going to happen after that. I agree. Yeah, I would say if you're, if you want to play with Irby, it's, it's fun to play with, and there's some interesting groups,

you know, there are actually quite a few different social groups on Erba that, that are very different from another. So the idea that we're going to foster the development of subcultures has been empirically true so far. But then I think the thing that's really exciting about Erbit right now is if you're a developer, because this is a really good time to get involved and play with it, right? Build something. Take a weekend. Maybe it takes a week, if you've got to learn the programming language first.

Maybe it takes a little longer, but it's not that bad. And then learn your way around the system, build an app, and you can build it and publish it. And it's really easy. and then talk to people about it. And you might get some nice positive feedback, make it actual users very quickly. And there are all kinds of apps that should be written for a crypto integration.

So if you're coming from crypto especially, and that's strategically interesting for us at Erbit. And you may be able to get a grant from the Erbit Foundation to work on something like that. We have a grants program. You should check out. Bounties. So there are a lot of different ways.

of getting involved. And also you can get on the network and talk to people about what they, you know, what they want to build, work with some other people. So yeah, I think it's a really fun time to get involved with Erbit because it's at the point where the, you know, app development and app distribution is somewhat mature. It's mature enough that it's not a huge pain. And then also, but despite that, it's still very young. And so there's still a lot of low-hanging fruit that needs to be written that's not too,

not too hard to do. Cool. Well, thanks so much, Ted, and thanks so much, Gary. It was a really pleasure to have you guys on, and there was a pleasure. I think this was a nice, a nice other introduction about Erbit, because I think often you need quite a few introductions about Erbit, so I'm pretty sure people will enjoy this. And I'm also sure this is not going to be the last time.

We're going to talk about Erbit, but there's going to be a lot more to come. And I think the next year, the next two years, and what's ahead. It's going to be incredibly exciting. So I can't wait to see how it all unfolds. And yeah, also so grateful for you, Ted, for doing all this work in trying to get this platform to maturity. Thank you. Really enjoyed this.

All right. And thanks to our listeners for tuning in. Let us know how you like to show on Twitter or leave us review if you want to support us. And then we look forward to being back next week. Thank you for joining us on this week's episode. We release new episodes. every week. You can find and subscribe to the show on iTunes, Spotify, YouTube, SoundCloud,

or wherever you listen to podcasts. And if you have a Google Home or Alexa device, you can tell it to listen to the latest episode of the Epicenter podcast. Go to epicenter.tv slash subscribe for a full list of places where you can watch and listen. And while you're there, be sure to sign up for the newsletter, so you get new episodes in your inbox as they're released. If you want to interact with us, guests, or other podcast listeners, you can follow us on Twitter. And please leave us a review on iTunes. It helps people find the show and we're always happy to read them. So thanks so much and we look forward to being back next week.