WEBVTT 00:00.000 --> 00:12.760 Now we welcome Eric, who we talk about seed hypermedia, the future of digital sovereignty. 00:12.760 --> 00:14.760 Let's welcome him. 00:14.760 --> 00:15.760 Hello. 00:15.760 --> 00:23.760 Okay, so my name is Eric, I'm on the founding team of seed hypermedia and today I want 00:23.760 --> 00:27.880 to talk to you about the future of the web. 00:27.880 --> 00:33.840 Because on our team we love the web so much, but as you know today, and I hope you do 00:33.840 --> 00:34.840 too, right? 00:34.840 --> 00:36.200 Does anyone here use the web? 00:36.200 --> 00:37.200 I think so. 00:37.200 --> 00:39.320 Okay, that's a few of you. 00:39.320 --> 00:44.200 The web today sadly looks like this for many people and for some of us as well, I mean 00:44.200 --> 00:49.640 on a practical basis, I'm using all of these, and I think this is really depressing that 00:49.640 --> 00:57.840 we have these companies in between us, moderating what content we're posting, and 00:58.320 --> 01:00.920 controlling our minds with the news feed. 01:00.920 --> 01:06.440 Basically, I believe we're living through the dark ages of the web, and it's not just 01:06.440 --> 01:12.920 the centralization, there's a lot of fundamental features that we are lacking in the web. 01:12.920 --> 01:15.160 So let's quickly talk about them. 01:15.160 --> 01:19.200 Have you ever clicked a link and nothing was there? 01:19.200 --> 01:24.760 Dead links, broken links, I think this is an incredibly detrimental to human knowledge. 01:24.760 --> 01:28.760 This is such a disaster. 01:28.760 --> 01:34.240 Also the web has no built-in archival system, I'm sure you're aware we rely on one organization 01:34.240 --> 01:40.640 as great as they might be to basically back up the web, and that's also the only way that 01:40.640 --> 01:44.160 we have on the web to know what was there yesterday. 01:44.160 --> 01:48.920 There's no built-in version history on the web, and so we use the way back machine, and 01:48.920 --> 01:53.440 I love the engine at archive, but we shouldn't be putting the world on their shoulders 01:53.440 --> 01:57.920 to back up the web and provide version history for us. 01:57.920 --> 02:05.120 The web doesn't have any unified identity system, so unless you want to log in with Facebook, 02:05.120 --> 02:12.000 you can't take one identity from one website and go use it to do something on another website. 02:12.000 --> 02:17.280 Also the web has a lot of friction if you want to publish something on the web. 02:17.280 --> 02:23.000 This is why I think people are using social media so much, because from your phones, you 02:23.000 --> 02:29.240 can't really spin up a server and FTP into a server and write HTML, so we have to rely 02:29.240 --> 02:35.320 on all these services that add so much friction to putting something online. 02:35.320 --> 02:40.320 Also when the web was founded, the hypertext community noticed that it didn't have any 02:40.320 --> 02:46.200 backlinks, so you couldn't see who is linking to you, and they told Tim Berners 02:46.200 --> 02:52.120 Lee, this web thing is never going to succeed, and because it doesn't have backlinks, 02:52.120 --> 02:57.960 and also links, you have broken, they were criticizing him for that on day one, and they said 02:57.960 --> 03:02.040 it's never going to succeed, and they were right about those things, but they were wrong 03:02.040 --> 03:05.720 when they said it would never succeed. 03:05.720 --> 03:13.160 Also the web, which doesn't have a trust system, and I think this is super, super problematic, 03:13.240 --> 03:19.240 because when you are interacting with somebody on the web, it's kind of like the wild west. 03:19.240 --> 03:25.160 You don't know who you're interacting with, and this has led to just an extreme proliferation 03:25.160 --> 03:28.720 of slop and spam. 03:28.720 --> 03:34.600 So for those reasons, I think we are enduring the dark age of the web, and I think humanity 03:34.600 --> 03:39.960 deserves a web where we connect directly with each other, and I think that people in this 03:40.040 --> 03:46.840 room and those watching have the opportunity to come together and become the inflection point, 03:46.840 --> 03:52.520 so that we can move away from the dark ages and towards the renaissance of the web. 03:52.520 --> 03:54.520 But how might we do that? 03:54.520 --> 04:00.720 I think there's a few ingredients, and for one, let's start with a foundational technology. 04:00.720 --> 04:02.800 This is why we're developing our protocol. 04:02.800 --> 04:08.600 We call it the hypermedia protocol, it's quite an ambitious name, I know, but we're going 04:08.600 --> 04:14.840 to stick to it, and so the hypermedia technology, hypermedia protocol is built on some really robust 04:14.840 --> 04:22.840 foundations of lip p to p and IPFS, so lip p to p is our peer-to-peer networking stack, and IPFS 04:22.840 --> 04:29.400 is what powers the content to add dressing of our system, content to add dressing, I'm sure 04:29.400 --> 04:34.680 most of the people in this room know, is a really cool concept, because it allows you to take 04:34.760 --> 04:38.840 a big chunk of data and identify it by its hash. 04:38.840 --> 04:44.600 So that means this data cannot change, which is both a feature and a future problem that we will address, 04:45.400 --> 04:50.600 but most importantly, this means that you can trust data from anyone, because if you ask for some 04:50.600 --> 04:57.240 data by its CID, they give it to you, you can validate it, so now you can get your data from anyone, 04:57.240 --> 05:03.240 and what that creates is an open network of archivists, so suddenly we no longer depend on 05:03.240 --> 05:08.680 one organization to archive content, as long as we keep track of some CIDs along with some future 05:08.680 --> 05:14.760 techniques. So we're going to develop a decentralized sink engine so that way we can handle mutations 05:14.760 --> 05:20.520 in our system. It's, you know, we've been talking about sink engines all day, but oftentimes 05:20.520 --> 05:25.320 your local first application is going to be talking to one server. Now we're going to start talking 05:25.320 --> 05:31.400 to everybody, and we'll see how we can manage that. At a very simple level, let's take a look 05:31.560 --> 05:38.120 at one node and note that we have some immutable chunks of data in here, which are our document 05:38.120 --> 05:43.480 operations. So it's describing how a document is changing over time. I'm going to move fast because 05:43.480 --> 05:48.360 I think everyone here knows where this is going. Basically, you can iterate over each of these to 05:48.360 --> 05:53.240 understand the current state of the document. This is Alice, she's making her own version with 05:53.240 --> 05:59.800 her own changes, and she arrives at her own version by iterating over each of these. Meanwhile, 05:59.880 --> 06:04.280 Bob is on the other side of the world. Bob is on a boat we like to say on our team, and so they're 06:04.280 --> 06:10.920 not connected, and they're making concurrent changes on the same document. So how do we solve this? 06:10.920 --> 06:16.280 Well, when they come online, they exchange the data, deterministically sort each of these 06:16.280 --> 06:22.360 operation blobs, these change blobs, and then go ahead and evaluate them, and they've arrived at 06:22.360 --> 06:30.600 the same state. So this is our oversimplification of our CRDT, and how many CRDTs tend to work, 06:31.720 --> 06:35.320 and I don't have too much time to go into it, but I'm happy to tell you more about the details 06:35.320 --> 06:40.760 of our CRDT later. So the other thing is that these two might not be talking directly to each other. 06:40.760 --> 06:44.040 Those blobs that are exchanged might have gone through a bunch of intermediaries. 06:45.240 --> 06:50.120 So how do you know, how does Alice know that those were Bob's changes? Well, of course, 06:50.200 --> 06:55.720 everything is going to be signed. So that means your private key is your identity, 06:56.920 --> 07:03.320 and what we've started off with is we started off with allowing you to create a private key 07:03.320 --> 07:06.840 and get it as a 12-word nemonic that you are then responsible for saving 07:07.880 --> 07:13.960 into your password manager, or writing down or whatever. It's your responsibility to keep it safe. 07:13.960 --> 07:18.120 But what we found in practice, trying to do this with real users like normal people, 07:18.760 --> 07:24.360 is that no one can save 12-words. They will forget to. They'll click the checkbox that says, 07:24.360 --> 07:28.840 yes, I've saved my 12-words, and then the next day they're going to say, I don't have them, 07:29.560 --> 07:34.520 and we get very frustrated. So we're adopting an approach, we haven't finished this yet, but we're 07:34.520 --> 07:42.200 working on it right now, where we store the encrypted keys on a remote server by taking a few keys, 07:42.200 --> 07:48.200 one of many keys into a vault, encrypting and saving them on identity server. And this is how 07:48.200 --> 07:54.600 password managers keep your identity safe. This is again an oversimplification, but I'm happy to 07:54.600 --> 08:00.520 tell you details later, and really I'm not the right person to be giving that talk. But this is just 08:00.520 --> 08:05.320 the tip of the iceberg for the foundational technologies that I think are necessary to build a better 08:05.320 --> 08:12.760 web. What else do we need? Well, I think we need really accessible interfaces. We need to make it 08:12.760 --> 08:18.200 super easy for normal people to get on to this thing, because otherwise we're just going to be 08:18.200 --> 08:21.960 living in our own world where all of the gigs are talking to each other using some cool new 08:21.960 --> 08:29.320 technology. So of course, we're starting with a local first desktop application. It looks like this. 08:30.280 --> 08:36.280 It acts as a browser. It acts as a publishing tool with a wizzy wig editor. It acts as an archival 08:36.280 --> 08:44.040 tool, and because it's a peer-to-peer node, it also acts like a server. But a lot of people don't 08:44.040 --> 08:48.680 use their desktop very often, or they don't want to install a new application, or they don't trust 08:48.680 --> 08:54.920 us yet. So we're developing a web application as well. And in the web application, the peer-to-peer 08:54.920 --> 09:00.120 node is running on the back end. Anybody can self-host this, of course. And it's used to power 09:00.760 --> 09:07.240 phones and lighter weight environments like Chromebooks. So it's eventually going to have 09:07.240 --> 09:11.160 basically all the same features that the desktop software has, but you're of course going to be 09:11.160 --> 09:17.560 happening to trust this server that you're talking to, trust this domain. Okay, so accessible interfaces. 09:17.560 --> 09:24.120 We need really rich content for you to be publishing. So we have a block-note like editor, where I 09:24.120 --> 09:31.160 sorry, we have a notion like editor powered by block-note. It's got all the different types of things 09:31.160 --> 09:37.320 that you might want, and we're hoping to make it accessible in the near future so that you can 09:37.320 --> 09:44.200 add new types of content into your documents that are sort of custom to you. So what else? What else 09:44.200 --> 09:49.800 do we need for humanity's information system? We need to be able to make really deep and 09:49.800 --> 09:58.040 precise references. So here I'm copying the exact version of a certain document. I think that's 09:58.040 --> 10:03.080 really important that you can be able to reference the exact version, snapshot it in time, 10:03.080 --> 10:10.440 referenced through the CID, the IPFS CID. Also, you can drill into a document however you'd like. 10:10.440 --> 10:16.920 You can reference a single paragraph. You can reference a section of a document because every document 10:16.920 --> 10:23.000 is composed as a hierarchy. And you can even reference just a string, just a section of a paragraph, 10:23.720 --> 10:27.960 a text range. And when you do this, you really should be referencing the exact version so that 10:27.960 --> 10:36.200 way your selection is not changing out from under you. What else do you need? Well, one thing I 10:36.200 --> 10:41.960 didn't mention is that the web doesn't really have the ability to embed part of one page into another. 10:41.960 --> 10:46.680 The web has eye frames, but this is not really what we're talking about. You might want to take 10:46.760 --> 10:51.080 part of somebody else's content and put it into your content because they did a good job, 10:51.080 --> 10:56.840 and you want to reference them. And when you make that reference, by the way, you're copying it 10:56.840 --> 11:02.520 into your server, but you're not stealing their content because you're not stealing attribution. 11:02.520 --> 11:08.760 It's all been signed by them. So this allows you to solve this problem of, do I link to the 11:08.760 --> 11:14.760 website where a link can be fragile or broken or do I copy it, which is kind of like plagiarism? 11:14.840 --> 11:20.840 This is neither. And this is archiving the content and making it so that their their page, 11:20.840 --> 11:27.400 their document is backed up just one step further and can be available to anyone who's 11:27.400 --> 11:33.800 browsing your site now as well. So we also support collaboration as I was alluding to with Bob 11:33.800 --> 11:39.960 Analysis and Bob being on a boat. And so when you add a collaborator, you're creating a certificate 11:39.960 --> 11:45.160 capability. So basically if Bob is inviting Alice, she's signing something that's saying 11:45.160 --> 11:50.280 Alice is invited to collaborate on a document. And this works a lot like you can if you're familiar 11:50.280 --> 11:57.720 with those. And we're going to go into a lot more our capabilities are pretty simple right now, 11:57.720 --> 12:01.720 but we do have functioning document collaboration. It's one of the most important things 12:01.720 --> 12:10.520 that we're going to save the web. Next up, we are not going to rely on social media as the 12:10.520 --> 12:17.160 layer of commentary anymore. Conversations are a first class citizen, not an afterthought. You can 12:17.160 --> 12:24.760 comment on anything, comment on documents, comment on other comments. You can cite anything from 12:24.760 --> 12:29.560 your comments and your comments can be cited using all of those same sophisticated techniques that 12:29.640 --> 12:36.920 we talked about earlier. Citations, which are basically backlinks, is you can see when somebody's 12:36.920 --> 12:44.280 referencing your documents. This is really, really important to create a sort of thriving knowledge 12:44.280 --> 12:48.360 system on the web. You shouldn't have to go to Google to see who's been linking to you. 12:50.600 --> 12:55.000 And we also want to make it really easy to publish. Even if you don't have your own server, 12:55.080 --> 13:01.080 you should be able to get on to the web. So we are running a free service at hyper.media slash 13:01.080 --> 13:09.160 hm slash public key, basically, where your peer to peer content can get on to the web with one 13:09.160 --> 13:15.880 click. So you use our desktop application. The desktop application talks to the A gateway server, 13:15.880 --> 13:20.680 which by default is our gateway server because we want to make it super easy by default to get on to the 13:20.680 --> 13:29.080 web. And there you go, one quick publishing from a desktop application, but keeping your sovereignty 13:29.080 --> 13:35.800 because the key is never leaving your device on encrypted. Okay, we also want to have really 13:35.800 --> 13:41.160 beautiful URLs. This is actually one of the coolest features of the web is that you can see 13:41.160 --> 13:48.440 what you're going to get before you click it. And this is a bit of a fragile URL because 13:48.520 --> 13:55.800 Eric.vescenty.net, this is a domain that I'm paying for right now because I'm alive. But in the 13:55.800 --> 14:01.800 chance that I die one day, it's possible that I will die one day. And I think that there should 14:01.800 --> 14:08.120 be a way to access my content anyways. So basically, my public key will allow you to address 14:08.120 --> 14:13.240 this anyways. And you can talk with your friends to sort of figure out, oh, this domain is gone, 14:13.320 --> 14:19.080 but it used to point to this public key. So eventually we can add a layer of what we call a 14:19.080 --> 14:27.320 distributed link base. This is just a term that we're using to basically resolve this into a 14:29.160 --> 14:34.120 hypermedia URL, which is clear to fear. Hey, distributed link base, let's go. 14:36.120 --> 14:41.560 So by the way, well, my domain is online. There is a mechanism where you can just immediately 14:41.640 --> 14:46.120 resolve it to the hypermedia URL. So whenever you're referencing it from your content, 14:46.120 --> 14:52.200 like you can paste in and you can open the normal HTTPS URL, but it gets resolved to what we call 14:52.200 --> 14:59.480 a permanent URL, a peer-to-peer hypermedia URL, under the hood. So all of this is like really 14:59.480 --> 15:04.920 relevant for your like users, like normal people, but I'm in a room full of awesome developers. 15:04.920 --> 15:10.760 And I believe that we need to deliver really great APIs, SDKs, command line interfaces, and agents, 15:10.760 --> 15:15.800 and genetic interfaces, because this is also part of making an accessible technology. It needs 15:15.800 --> 15:24.360 to be extensible and accessible to everybody. Okay, I probably need to move faster. This is two out 15:24.360 --> 15:29.720 of the three ingredients that we need for a new web. But lastly, I believe we need to invest 15:29.720 --> 15:36.840 in the web of trust, because the peer-to-peer web is messy and dangerous. It's not as simple 15:36.920 --> 15:43.080 as Bob and Alice and Bob being on a boat, because, you know, there's a lot of us out here. 15:43.080 --> 15:48.680 I don't even know which one of these is Bob and Alice. I didn't keep track. And a lot of these people 15:48.680 --> 15:53.320 might not have your best interest in heart. Like they might be running faulty software, 15:53.320 --> 15:58.920 they might be spammers, they might be attackers. So what's really important is to hold on to your 15:58.920 --> 16:03.960 identity and be able to associate it properly with people in the real world, to be able to associate 16:04.120 --> 16:09.560 it identities. As we talked about, we're supporting domain names. We're not going to be selling 16:09.560 --> 16:16.920 a token for you to get a name. You can resolve a domain name to a public key using HTTPS and 16:16.920 --> 16:20.600 pretty soon we're going to allow you to just use a DNS server because that's simpler for some 16:20.600 --> 16:26.840 people. And you can check the signed data that your public key is signing to see what domain 16:26.840 --> 16:33.080 is resolving to. So we have this really strong link when it works, but sometimes the domain might 16:33.240 --> 16:37.640 change, sometimes the key might change. So we're going to need to build software that makes this 16:37.640 --> 16:42.440 a little bit more sophisticated in the future. And some of this complexity will actually need to 16:42.440 --> 16:47.720 be exposed to users, which is unfortunate, but it's just a reality of having a stable identity 16:47.720 --> 16:53.960 that you properly own. And we're going to build upon more trust systems, not just domain names, 16:53.960 --> 16:58.600 but we're going to make it easy so that your personal relationships can exchange public keys 16:58.680 --> 17:04.120 basically face to face. And of course we're going to link your keys to social media accounts 17:04.840 --> 17:10.200 and basically any other system that you're interested in. If you know what key base is or was, 17:11.240 --> 17:20.440 this is that and we're just trying to do it again. So we need to have signed social validations 17:20.440 --> 17:27.480 so that one key is providing trust metadata about other keys. Usually it's saying like, oh I'm 17:27.560 --> 17:33.000 following you, I like you. It's much like following on Twitter, but Elon Musk doesn't need to be 17:33.000 --> 17:41.080 involved. And what we're going to do is we're going to form a web of trust. So that way we can connect 17:41.080 --> 17:47.800 with each other in a reliable way. I might trust a couple of you and you might trust each other, 17:47.800 --> 17:55.080 and within a few hops we know that we're in a really good space. So these are the ingredients 17:55.080 --> 18:00.520 that we need to build a better web. I'm really excited about this if you can't tell. 18:02.360 --> 18:08.360 If you're interested in our software you can check out c.hyper.media or get hub where a patchy too. 18:10.360 --> 18:15.160 Yeah and we're also opening this up as a protocol so that you don't even need to use our software 18:15.160 --> 18:21.160 if you want to communicate in this way. Although it's a little bit underbaked and under documented, 18:21.240 --> 18:24.920 but get in touch with us if you're interested and we're going to support you however we can. 18:26.600 --> 18:33.880 But before we finish, I do have to be honest with you because when I tell people about this project, 18:35.000 --> 18:41.480 they seem excited, but also I can tell that they've had their hearts broken before by similar 18:41.480 --> 18:47.800 dreams that did not pan out. People will say, this has been tried so many times before, 18:47.960 --> 18:55.640 but what they're really saying is that this is impossible. And sometimes this gets to me because 18:55.640 --> 19:01.160 I've also had my heart broken when it hasn't worked out in the past and people have had these 19:01.160 --> 19:07.240 really ambitious dreams and there are so many inspirational projects that haven't ignited the 19:07.240 --> 19:13.480 revolution that we need to improve the web. These are some of our heroes. They're the creators of 19:13.480 --> 19:18.440 the Hypertext movement and they spent a lifetime trying to implement these ideas in the modern world. 19:21.480 --> 19:26.440 We've learned so much from so many initiatives and I'm going to show a few of them that have inspired 19:26.440 --> 19:33.640 us in no particular order, but they haven't all survived. Some have faded away. Some have shut 19:33.640 --> 19:40.040 down or sold to a big company. Some of these projects are great but they have a very focused scope 19:40.920 --> 19:46.520 and they aren't trying to solve the biggest issues that we face in the web. And some of these projects 19:46.520 --> 19:51.960 live on but regular people haven't managed to adopt them. So when it comes to building a better 19:51.960 --> 19:59.000 web with mainstream adoption, it does seem impossible and it will seem impossible until we've done it. 20:00.600 --> 20:06.840 So I believe that it starts off with a few dreamers who have a vision for a future that everybody 20:06.920 --> 20:16.760 else sees as crazy. But before you know it, the concept is proven and what I find amazing is that 20:16.760 --> 20:23.960 in a very short period of time it becomes integrated into society so deeply that most people don't 20:23.960 --> 20:35.160 even think about it. So I believe we're including the most important ingredients but I don't believe 20:35.240 --> 20:41.960 that's enough. Our team has been working on this for five years and we refuse to give up, 20:41.960 --> 20:50.760 we refuse to pivot and we refuse to narrow our goals and even that is not enough. We need community 20:50.760 --> 20:57.240 leaders, knowledge managers and developers like yourself because together I believe we have the 20:57.240 --> 21:02.440 opportunity to build a better network for humanity. So what do you say? Do you think we can build a 21:02.520 --> 21:04.520 better web? 21:19.960 --> 21:21.960 That's quite a right. 21:23.800 --> 21:29.480 We reached the end of the day of the room. Do you say she wants to come with me to say goodbye to everyone?