 You're so lucky because we have with us on this panel a security expert, a startup technologist who was also once an activist, and a technologist who is now a policy fellow, a policy technologist. So we have it all covered on this panel. I think it's all about power. When you think about Corey talking about monopolies and we're talking about centralization. And that consolidates power. Smaller players, the ability to do new things. It gives us the ability to make choices between who will provide us with resources and services. And it all works together. It all interoperates. That is in a sense the original power of the internet. That's what email is. I mean now it's becoming centralized. Like there are five email providers, but there used to be thousands, tens of thousands. And no matter who you got your email from, no matter how you read your email, anyone else can email you. That is decentralization. And the more we have those types of protocols on the internet, the more we reduce the power of the large tech monopolies and the better we are in every respect. So when I think of a decentralized web, I think of the original web. And you will find through history, this is an accordion of centralization and decentralization. As different areas of the stack get in concentrated power and then society does something to them and then decentralizes, then decentralizes. And now we're in an area of, I think, great centralization. I think we can see it in the lack of interesting things. Platforms that you just mentioned take a form that we describe as federated. Which basically just means that it's just as if, so to take the short form social media example, Twitter for example, it's just as if there were a whole lot of different Twitters. But anybody from one Twitter could follow someone from another Twitter. And just by saying their name at a domain. And if that sounds familiar, it's because that's how email works. And email is a federated system. You can send an email from your university.edu account to your homes.gmail.com account. Sorry for saying that. And again, presuming that people doesn't deem you to be as famer in some way or shape or form that is only known to their unique algorithms, the email will go through and it will work. So the Fediverse, as it's described, is exactly just that. Except we use short form social media instead. And so you can go out and pick a Fediverse server that you want to join, which everybody seems incredibly intimidated by, but it's actually just as simple as choosing an email provider and everybody did that. So it's going to be okay. And then you just start following whoever you want to follow. And the only real difference from Twitter is that you just have to put an app domain name after the username to fully get to the person you're going for. And the other really, and what I think is the coolest part of the Fediverse, is that I've been talking about it as sort of a Twitter analog, but it isn't. It's not a Twitter analog. It's a protocol. And what that means is that anybody can build anything that is sort of even vaguely social media-esque on top of what's called activity code, which is the name for the protocol that underlies Mastodon and all the rest of these services. And it will be able to talk to the entirety of the rest of the Fediverse. This is what I think is really amazing and it's really hard to get people to grasp until they actually see it in action. So see if you can picture this, right? So you've got your favorite YouTube content creator, I guess is the term we're using these days. But you want to follow them on YouTube? You can only follow them on YouTube. That's the way it goes. On the Fediverse, there's a YouTube equivalent. It's called PeerTube. There's actually probably more than one, but PeerTube is the biggest one. And it's exactly the same. It's just YouTube. You can upload videos, you can comment videos, you can like videos, you can do whatever. But what's really cool about it is that from your Mastodon account, you can follow your favorite creator on their PeerTube account. And the way around, if they want to, if this PeerTube wants to follow you on Mastodon, they can do that. And when your favorite creator posts a new post, it shows up in your Twitter-esque timeline on Mastodon. That's pretty cool. That's handy, right? Now I'm going to look at one place to see when my favorite content creators post something new. But it goes deeper than that. If I like a video, if I click the like button in my Mastodon timeline of a video that somebody posted, that like shows up on PeerTube. If I reply to it as if I just replied to a tweet, it shows up as a comment on PeerTube. All of these things are connected to each other. And that's the like-bulb moment for a lot of people that said, hey, you know what, wait, I don't need the permission of somebody else. I just implement this protocol. I just write some code that, or frankly use somebody else's code that they've already written to interact with ActivityBub. And I get all of this for free. And it works across all of these platforms in a way that they make sense depending on the platform. So, you know, a reply on a Mastodon looks like a comment on PeerTube and so on and so forth. And there's, you know, a bunch of other examples too. There's an Instagram equivalent called PixelPet. There is a Goodreads equivalent called Bookworm. There are others I'm going into now. But that's sort of permissionless kind of, hey, you know what, I'm just going to implement the protocol and we're going to interact with the whole rest of the world that's out there without having to say, hey, please, Mr. Real and Musk, can I use your API? And that is really the incredible power of the fediverse and of the protocol itself. So, the word I want you to learn here is generative. And what these open systems are, are generative. That you can do a new thing with the protocol without anyone's permission. So, I mean, I sort of baked off solid before, but we have it a little bit out now. Solid is a personal data store. The idea is that right now, your data is stored by the company with which you create it. Your Fitbit data is stored by Fitbit. Your email is with your email provider. Your tweets are with Twitter. And you can't put them together. And you can't create something that uses them in an interesting way because they are separate silos and they're all controlled and, you know, very, very securely by the companies that add on them. If you had this kind of federated system or a federated system of your data, like solid, you can invent something new. So, like some app that marries your Fitbit data with your refrigerator data and your thermostat and, you know, what your doctor's data feeders made that up. But there are a gazillion things you can just make up. And that's the value of a distributed system that people can just make it up and make up something new and different and it'll be cool. And you've never thought of it before until you see it. Important controversies here are an important kind of nuance to these things that I think matter a lot. And I think one way to get out is to talk about how some of the picture that we just heard about, the federated, the sort of Fediverse, all these sort of federated sites playing together, breaks down in reality and how also I think the sort of idea of the personal data store breaks down and how I think we do need not a lot of new tech, or maybe not any new tech, but we need to get better at applying some of the tech that we have because there are some technical challenges here that aren't just policy challenges. So the way I'm... So first of all, I'll say how many people here have their own server? So not that many, right? Not a majority. We're in a very specialized context here, right? So you can imagine that the incidents of having one's own server here is higher than it typically would be. So most people don't have their own servers. My former organization, to solve the problem that I described before, actually did end up using Matrix, and it's great because we have our own server. We have a CTO who can maintain it, who can run it, and having that is necessary to getting the security properties or most of the security properties in Matrix because if you don't have that, they break down someone spectacularly, or at least until recently they did. And so there's a way in which, first of all, to access the full power of these things, you need your own server. One thing that Corey said earlier about Macedon is not actually true, as I understand it. You cannot just move your account from one service to another when your server had been table-flips, and it's like, I don't want to do this anymore. You can't just get out. They have to let you out because you don't have anything that you can just take somewhere else. You have to almost set up an email forward. And so that's really, really different than the ability to leave when you want. They still have the ability to make your 300,000 followers go away. So to look at it from that perspective, as Fight for the Future, or an activist organization, how do we preserve more as a political campaign? How do you preserve your access to your audience built up over time? Well, you might say, you don't want to just use some Rando server, obviously. And you don't want to use a big company server because you can be censored by them if you're in conflict with them. So you run your own server. Okay, great. But what server are all the people who follow you using? Right? And if they're all using G, because that's the problem that exists for email, we can use our own email services. We can switch, as Bruce said, from email to email. But if all of our subscribers are using Gmail, and Gmail wants to block us, then we're dependent on Gmail for that relationship. It's two-sided. And if you actually look at how, even the systems like, and I've talked to the Blue Sky founder, Jay Graver, who actually used to work at Fight for the Future, was a campaigner with us before they learned to code, and now they're making decentralized Twitter. But I talked to Jay Graver about how this, like, games out, or how it plays out in terms of the game theory of it. And it's almost like nation states. But there is a dynamic where you end up flocking to whichever server is biggest. So the dynamic that Bruce talks about a lot in his work too, like feudal security. Anyway, the way it plays out is if you're a Fight for the Future and you want to maintain access to your audience, the best thing you can do is be on a server that has all the Taylor Swift's of the world and all the people that nobody is ever going to block. And so you're going to flock to the biggest service that is like that, and probably that biggest service is going to have something pretty similar to the rule of law, like a very transparent, consistent set of criteria that if you meet these criteria, we are not going to block you. And then you can be sure that your audience will have access to that because no one is going to want to run a service on their side that doesn't let them see Taylor Swift or the sort of biggest folks in the world. And then you get into what happens. What if somebody is really popular with one group and really anti-popular with another group and you can get fissures, right? Almost like sanctions where different services end up kind of blocking each other in a retaliatory way. It gets weird. But the tendency is for there to be a few big operators and that is in fact what has happened with email, right? Most of us email, you can theoretically run your own server, most people don't. And so the reason and motivation for my project Quiet is that I'm really interested in how we can get past that dynamic where as it seems has happened with email, even when we have open protocols, over time it tends to people blocking to whichever party runs the most secure, reliable, and more or less okay for most people's service and kind of getting locked in there. And I think we do need some tech for that because your laptop is not the same as a server. I'm sure you can have all your data on your laptop with your laptops going online and offline every time you close it. Switching IP addresses every time you go from your house to your coffee shop. And so what the software and what the tech needs to do in order to deal with that and create a consistent user experience that feels like you're using Slack, when everybody's devices are going on and offline in that way is actually quite nuanced and interesting. And it can be done for the team collaboration case right now. And that's why I'm tackling this and making Slack and not trying to make Twitter in this peer-to-peer way. To make Twitter in this totally peer-to-peer way on devices that are going on and offline all the time, that's still an interesting academic research area in the field of peer science. That's a big question. But it would be great if we could do it and quickly to close, I think, here's the reason why. There's this long history of open source getting us out of jams when it comes to monopolies and competition in the history of tech. Like, it happened with Microsoft in the 90s. The open source web browser, the open source lamp, SAP Linux, Apache, MySQL, and PHP let people make applications to distribute them directly to users, kind of going around doing an end-run around Microsoft's electrification, maybe, of Microsoft's operating system and get access to build applications. That's where the entire field of big Silicon Valley and tech companies started, essentially, Google, Amazon, or things like Dropbox. They all came from this ability to just build and distribute directly to users via the web. And all of that was based on open source software. All the products we used are at, like, 90 or 99% open source, depending on how you measure them. And the companies that build these products, their giant software teams just work on this tiny sliver at the top to differentiate the stuff that's specific to the thing that they're building. But they collaborate on this huge open source stack underneath them. The problem is what's happened is that, right now, that open source software is sort of free and accessible to everyone in practice. Like, you can go and look at the code that drives these big services, but the ability to package it all up and run it at scale in a way that's reliable is a highly specialized and intensive set of skills that's very locked up in these organizations that just happened to get there first and have to solve those problems first. Now they rent it out to the rest of us, but for an increasingly high toll. And the way to get around that is to try to wrap up some of this operational capacity that these companies supported internally and package that up with the open source code itself. So if you... The contrast between Matrix and Quiet is that if you had the code for Matrix, you would still have to find someone to run your server before you were doing it yourself. Whereas... So the operational capacity does not come with the code. Whereas if I give you Quiet, you don't need anything else. It just works and it will connect to the devices of different people in your team. The operational capacity comes with the code. And so that I think is, in a sense, a response to Bruce, that there is some text up that we need that can bring some of this magic of open source software and generativeness to the problems we're facing with Big Tech if we solve a couple of specific problems. We'll take a quick second to quibble with your characterization and say that part of the reason that not everything is scaling and dominating is because not everybody wants to scale and dominate everything. It's actually a requirement for humankind to dominate everything. It's okay to have small things every once in a while. To answer the actual question that you're asking. So the fundamental issue is that all of the problems that you might encounter trying to run a centralized service, say social media service, absolutely exist in the decentralized world. You have to solve all of them. You have to solve some new ones sometimes, actually. And they don't get any easier necessarily just because you're running a decentralized service. You still have to tackle all the same problems. One problem that you just sort of alluded to is that oftentimes you don't have nearly the resources that a huge centralized service has. Either because the customers are paying for it or because you're advertising to them. And so we're seeing kind of play out now in real time this sort of these two sweet spots in the Fediverse. At the large end where you can sort of approximate being a small Twitter, right? So if you're like Macedona Social where most people are, you have some resources or if you can get people to kick in $45 a month, which, yes, $45 is not much, but if you get all of your users to do it, maybe you've got something to work with. And at the other end, you've got this sweet spot of hey, I'm running a friends and family server for 30 to 50 of my best friends. And frankly, I just kind of trust them not to be doing stupid stuff on my server. And so I don't have to expend a lot of resources in running it or administrating it or whatever. And then you have all of the stuff in the middle where you don't have the resources, but you also can't trust everybody on the server. And that's where we're seeing the people table flipping. The admins are saying, you know, this is not worth it. This is not what I signed up for. This is not fun. I'm getting yelled at basically by everybody for everything all of the time. And nobody's giving me any money for this. And so, you know, this is one of the forces that tends to trend life generally, things back towards centralization, right? Is that centralization is efficiency. Centralization is greater resources. And combating that and trying to drive a wedge away from that is both finally important and also very difficult because our capitalist system is set up to drive for efficiency. And so pushing against that is really hard. And you end up, but you do have to deal with all of these. The same problems, privacy, security, harassment, management, moderation, et cetera, et cetera. All of these problems still exist in the Federalist. I mean, it's all fun and games. I'm told, what are your friends' sons post national secrets? Just grew to his group. And suddenly the FBI is raiding everything and then ruining your day. Actually, I have a different quibble. Actually, that's a quibble. I think you explained the benefit. If you have your email hosted by somebody, maybe schnari.com is hosted somewhere, and they suddenly go out of business or get weird, I can move it. I can move it tomorrow. There's no problem. I just move it. That's the benefit of a distributed system. Now, you're right. We all could run our own email servers, but we don't. We likely pick someone. But it'd be great if there were 10,000 we could pick from. And so if you have a pod where you keep your data, using the solid protocol, you could run your own solid server. You're probably not. You're probably going to coast it somewhere. But you'll be able to move it. If you don't like the way Google is treating it, you go somewhere else. And hopefully, there will be tens of thousands. I think you're right. The dynamics of our markets push towards centralization. But we can deal with that as well if we choose to. We choose as a matter of policy to allow the big to grow bigger, because that's what the big like, and the big controlled lobbying gull is the control of the policy. But we can choose another path. I think there's a funny thing that happens in the tech policy space, like after the Snowden revelations, we find out the NSA is spying on everyone. And I don't know, others here probably saw this too, but there's this funny dynamic where all the tech people said, there's no way tech is ever going to solve this. We need policy solutions. We need some type of Congressional interaction. And then all the tech, all the policy and politics people said, there's no way policy is ever going to solve this. Like the national security narrative has a total grip on everything. We need tech interventions. And I think that happens here to some extent, right, where it's easier. And I think one fun thing for being someone who's like really deep and gets deep in tech and gets deep in politics is that the answer is sometimes both or neither or one or the other, depending, not always both though. Sometimes it really is like policy can only, you know, the privacy space, there's no tech intervention that can help you against light display readers, none that I know of or against satellites just watching you walk around. That's not a tech thing. That's a policy thing. On the other hand, there's no policy conversation we're ever going to have that's going to get us global private money because no government is going to allow that. So if that's what you want and think is important, you have to do it with tech. So I think there's like a thing here where in this case the tech intervention might be pretty good because yes, in general corporate America, we do let things get really big or the world does let things get really big. But in some parts of tech that hasn't been a problem, it's not a problem that Postgres is really big or that... It's a database for those of you who don't know. Yeah, sorry. So Postgres is a database that runs most of databaseing right now on the internet. And it's not a problem that Linux has gotten really big. Linux is another big fundamental piece of the stack or Ubuntu, that Ubuntu has gotten really big. It's the leading Linux distribution or was written until recently, but it's one of the big ones for servers. When these things get really big, it doesn't matter because they're open source, they're multi-stakeholder, they're points of collaboration across lots of different companies and organizations that have lots of different interests and needs. And in that collaboration around these shared license code, all these companies get what they want. They can be not just companies, but individuals and organizations. They can all be very productive, maximally productive. They can have the best database, the best distro that humanity can come up with at any given point without having to do much work to get that. And it doesn't matter how big it gets. The bigger it gets, the better. I mean, the bigger Wikipedia gets, like if it stays high quality, better. We all share in that. That's like a common resource. So if we can move some of these things from the space of the private and the control and then maintained by somebody into the space of this is code, this is non-scarce, infinitely copyable, common resource that we all share, the bigness isn't as much of a problem anymore. And then it doesn't matter that we're so bad at dealing with bigness, right? And so that is where tech can help. And specifically, one way we can do that is by making apps where you don't need to run your own server and where running your own server isn't just a theoretical thing you can do. It's just like kind of what everybody's doing, although you don't think of it as server. It's just the app is on your phone doing the thing. That's what's interesting about that line of tech is that we can create a new thing that no matter how big it gets, it will have the same maliciousness over time as Linux or as Postgres. These very, very neutral tools that become shared points of collaboration. Yeah, that certainly was a big part of it. And although I think, as Ross mentioned, abuse is also a part of it. So not just spam, but usage that some subset of users might consider legitimate uses, but some broader set of users consider important. That's also a piece of it. The big open networks case is really difficult. Yeah, to throw a bit of a wrench in here, one thing that came up was that whoever's dealing with it is in a position of not being paid to deal with it. And money came up as an important piece of this a couple of times in talking about it. Well, there's no money for this. There's no money for that. And these big companies have money, but these small operators don't. And it is notable that there is a piece of the decentralization space, a small piece, I would say, in terms of significance, but a big piece in terms of the amount of money going into it. Cryptocurrency, where there are building blocks for building money into these systems. And I don't think there are great examples for projects that have kind of threaded the needle in terms of building a cryptocurrency solution to problems so multifaceted as the kinds of problems that a master's on and then would have to encounter. But it's conceivable. You know, we have an open source programmable basis for building you get paid into these systems. So it could be you get paid if you do a really good job handling spam abuse. You get paid if you do a really good job creating a culture or creating a set of mores for a certain part of this community. There's also, we're seeing other, you know, I'm talking about the technical versus policy kind of solutions or we're seeing really interesting stuff happening, again, in the sort of master on hosting space with basically like user-owned cooperatives, where again, you know, a lot of servers are like, hey, yeah, if you could chip in a couple bucks to help maintain this stuff, that'd be great. Social.coop, which is a user-owned cooperative master on server, is like, in order to join, you must up, you need to, I don't even know how much it is, X number of dollars. And that goes toward running the server and maintain the server and paying people to do the things that keep the server in a nice place and so on and so forth. And if you don't need a blockchain for that, you just need a social construct as old as literally human beings, a cooperative, in order to make it happen. Although spam is an interesting example of the way it's built, leads to the tech. Back when there are 10,000, since the thousands of email servers, you had had anti-spam on your email client. I remember Udora used to do anti-spam on the client. And that's because that was the best place to do it. As you moved into a centralized model, the spam is being dealt with at Gmail, right, at Apple. In the middle, you remember those spam proxies that you would hire them and then they would clean your email pipe before it got to you. So there are different ways to implement this depending on the architecture of the overall system, how centralized versus decentralized it is. So I mean, if whatever it is running is decentralized, they will be decentralized solutions. And then it might be something running on your computer that cleans out the spam and hates speech whatever it is you don't want to see. And it might be something more centralized, but it really depends on what's being done and with a mix of centralized versus decentralized. Nothing is wholly decentralized. The very fact that you're going to buy a laptop by Dell or a compact or one of a dozen companies is a point of centralization. The chips, the different services, the operating systems, they're all, you know, there are different numbers of them and it's often not big numbers. So depending on where those centralizations are, you'll see different types of solutions. Let's talk about money a little bit because so far we've talked in very utopian terms, right? It doesn't take a lot of money. You can keep it small and bespoke. But the truth is that when we started this conversation, Ross and some of us in 2016, we thought it wouldn't scale because there wouldn't be money in it and you need money to scale. But what happened is cryptonomics happened, cryptocurrencies happened. And instead of these decentralized companies going to a VC and saying, okay, give me this amount of money and then the VC has a lot of say in how you monetize, how you grow, what your goals are. These decentralized protocols found that they could just offer an ICO and sell tokens and raise literally billions of dollars if you're Ethereum or your Filecoin or your Zcash. And they use those tokens to build their technology and in some cases to pay the open source community that builds that. Has anything we talked about done that? Linux didn't or actually didn't. Macedon didn't. Ethereum. Ethereum, Matrix, Filecoin have all. Does Matrix did? No. So I don't think that's the thing. That's true. Yeah, so nothing we talked about does that. But some of the stuff that we have talked about has these holes in it where no one's really sure what to do. Like there's this problem of if I put a video out into the world or I put a lot of video out into the world. Video is fun because it's just a lot of data. I put a lot of HD video out into the world. I want to make a new YouTube alternative and people are or you have a messaging app and it's using everyone's sending videos to each other. There's something we're like that does at a certain size that does become a scarce resource. Like you have to keep that video maybe forever. Maybe users probably want you to keep it up forever. You have to transcode it into a bunch of formats. There's different, there's computer requirements just to get it to show on all the different devices and there's always more devices. And you have to solve these questions of who does that work. Is it the work of the person receiving the feed? Is it the work of the person putting the video out into the world? And there is like scarcity there. And where there's scarcity, where there's some scarce resource where you can't just like copy things infinitely the way you can with code. Where there's like a computer using electricity using hard disk space. There's some balancing and interest you need to figure out between the different participants in the system. And in some cases there's enough just general goodwill and there's a small amount of data that it just all works out. And a good example of this is something like Tor where there's a network where people volunteer to run servers and it's available to the world. Everyone can use it for private browsing which works. But a Tor type network wouldn't be able to run YouTube and I don't think anyone here is arguing that it could. And it's not clear where that line is and whether you could have YouTube operate at the scale that YouTube operates at a global scale with a global ability to pay to maintain it. Because some people might be spending 45 bucks a year on a Macedon service but like not everyone in the world is going to do that. And at some point you might need some piece where you actually have to pay for that electricity and pay for that hard drive space. And what's cool about the cryptocurrency piece being built into this and what's useful for it is that it handles those cases, conversations. But I do think that there is a sliver of situations in which blockchain solves an actual problem that actual people have. I think if people are familiar with the Gardner hype cycle the problem is we hit the top of the blockchain type peak or whatever and it just so happened that that peak was also defrauding a massive number of people out of their hard-earned cash. And so now we're heading into the trough of wow that was really a poor idea. But I do think that eventually we will come out on the other side of that with a handful of really good uses for blockchain and hopefully only just those handful and not all of the massive fraud that went along with it. But there are some cases where this is a usual data structure and like you said, it's a community playing for some common sort of thing. I think identity management is actually another one. But yeah, we're unfortunately not there. We're still in the kill the fire mode. So I think to challenge the concept a little bit especially these days I'm not necessarily sure that everybody has to pick the same protocol as long as the protocols are open and we have a way to translate between them. And so like if some folks want to run Blue Sky and some folks want to run ActivityBub or AT protocol, sorry, which is the protocol underneath Blue Sky, I think we can probably find ways to make those talk to one another especially if they're sort of going out of the same general thing which appears to be social media exchange and post kind of stuff. But I think the point is still kind of taken we do need to sort of, it's good it's good to try a whole bunch of different things to see what works and what doesn't work but it's also good that at the end of the day everybody can sort of talk to one another. And we are, especially in the kind of the social media setting we're still in the kind of try a bunch of different things mode, right? So Blue Sky or the AT protocol has a couple of different things about it that are I think manifestly better than Macedon. I think Macedon has some things that are better than Blue Sky but trying to do things and seeing what sticks is a good thing at the end of the day but I think that as time goes on we'll start to see a coalescing around a couple of different protocols at the most that tick all the boxes and that everyone can kind of agree yeah we're going to, this is what we're going to use going forward. In a sense, protocol is the heart of internet coverage the internet works not because anybody forces anybody to do anything but if you don't follow the protocols you can't talk to anybody. So whether it's the ITF which is very distributed, bottom up they vote by humming, no joke Solidus W3C protocol World Wide Web Consortium and these are all tech organizations building protocols so that things can talk to things not really run by companies they're involved but their power is iffy depending on which one and this feels like a really decent model as most of us say the telephony protocols which are very much where governments pushing the protocols down to the telecom monopolies to the people. Internet I think is a better system. I think it's really hard, oh jump in quickly, it's super quick idea the one really hard thing in this space and actually I think this is the answer to the reason why none of this stuff is doing that well is that if you have to innovate on protocols and stack in order to build your product you're kind of like breaking the rules of how to make product. The start-up way to make product is you just cobble something together to see if it fits the needs of the people you're building for and try to get traction and then you kind of build it out from there and what's been great about the centralized stack is it's really easy to do that up to like 10,000 users or something and by the time you have 10,000 users you know you're on to something or you have an idea that you're on to something. The hard part about defining protocols up front is that we still don't know what the thing needs to do in order to attract the new product to be specific, needs to do in order to attract people away from the old products. It maybe isn't even enough that it has to just do the same thing the old products do just as well. It might need to do some new thing and figuring out what that is depending on what that is it changes what the protocol needs to do and what the stack needs to do and that interplay is really, really hard and so if we over define standards up front you could end up making it so that great you have a standard but no one wants to use your thing because there's some subtle difference between what the standard does and what people want and so what I'm trying to do is really let the work I'm doing be led by what I think people want and then work backwards to find the stack and then hopefully something will emerge that is some kind of consensus. So this is kind of secreted by design you know hopefully we get for new things I mean lots of organizations try to push this but the market doesn't reward that their market rewards you know fast and barely works not secure I can be secure I can be more secure than his protocol and I'm six months later and he wins so I'm not going to do that so it's a bunch of forces going on but you know I mean don't disthearally found us the internet they made smart choices given the hardware and software of the time I think what we may say I completely agree with Bruce like the internet that we have is the internet that we are going to have like no one's going to tear out of TCPIP and start over from the beginning but you know path dependence thousands of years the AIs what the hell is this thing what is it here but I think one interesting thing to sort of keep an eye on is this concept of like overlay overlay networks which basically you know underneath it is still using TCPIP because that's just we're going to be stuck with that forever I mean we can't even get IPv6 to work like we're not upending anything else but on top of that we are seeing some people starting to develop interesting networks on top of that that do have that are say encrypted from the get go or have more of like a mesh kind of framework than a spoke kind of framework so there is sort of some play happening there that may lead to interesting things that live on top of what we today call the internet but are not I don't think actually fully replace it at a base level