Building a Resilient Web for Decentralized Long-Term Content Storage

Dave Lockie:
You’re listening to Emerging Tech, part of the Open Web Conversations channel, which is an Open Channels FM production. I am your host today, Dave Lockie. I work at Automattic as a strategy wrangler. And on today’s show, I’ve got Sam from Arweave. Hi, Sam.

Sam Williams:
Hey, Dave. Great to be with you.

Dave Lockie:
So I guess like a little intro to yourself how people can find you and then tell us, Sam, what is Arweave?

Sam Williams:
Yeah, sure. So I’m Sam. I founded a decentralized storage network called Arweave. Basically the idea of Arweave is to build a permanent knowledge base for humanity that is shared between all people and not controlled by any one centralized party. It takes the idea of storing data inside blockchains, which actually, interestingly was the first use of a blockchain. Satoshi Nakamoto stored a record from or the headline from a newspaper in the first block of the network long before it was used for any kind of value transfer. But the problem was that and well, that is interesting because that information, that headline, is replicated in many, many places around the world now. It’s really very, very hard to believe that we will lose that information for hundreds of years from this point. We thought that was exciting when we were starting to build Arweave, but we saw that you couldn’t store large amounts of data inside. So essentially with Arweave, we’ve created a permanent decentralized web that lives inside a blockchain like structure where rather than people paying, well, the network paying for essentially expenditure of energy, it goes instead to paying for replication of the network’s data set. And then there’s an endowment inside the system which pays for that data storage over time. So that’s sort of what it is at the base layer. But what it looks like when people engage with it is a. Well, it looks like the normal web, except you can access it through many different entry points. And each of those entry points is independent. And so there is no centralized controller, no person that can come along and censor or change the data is open and neutral and transparent. Which I think might be interesting for the WordPress community because the thing that I see when I look at WordPress from the outside is, well, this is where the bastion of a decentralized or distributed Web still exists in the modern age. WordPress is this huge part of the Internet as it currently stands and is about the only place, I mean, yeah, it’s hard to think of any others really, where there are many different disparate websites that are run by different operators that don’t one way or another roll up to a single centralized party that is controlling it. Everything else has moved to, you know, the, the major web platforms essentially. And so, yeah, I think there’s a lot of sort of aligned ideas there and I’m excited for how Arweave can be used basically to replicate that data inside WordPress sites and make it first more permanent, but also more resilient. So it’s now not just dependent even potentially on that one sort of silo remaining online, the WordPress website itself, but could potentially be pulled down in other places and accessed through more decentralized means. Yeah, so that’s the sort of broad outline and about. I just got started building this because I saw that, you know, we live in a vastly different information economy than we did before the Internet. And yet the archiving technologies that we use are essentially ancient. They’re the same ideas. And so we felt that an Internet native archive was necessary and I equip my PhD doing distributed systems design in order to build it. And that was what, eight years ago? And it’s been incredible ride at this point. Now there’s 20 billion pieces of information inside Arweave and yeah, growing all the time.

Dave Lockie:
So 20 billion, that is impressive. So, yeah, like the, I guess just to recap and make sure I’ve got it right, if Bitcoin is a decentralized protocol for Internet native money, then Arweave is designed to. Well, it is the same idea. It’s a decentralized protocol, but for content storage.

Sam Williams:
Information storage, yeah, decentralized protocol for exactly trustless information storage where information is published and is in the public domain outside the control of anyone.

Dave Lockie:
So one of the challenges of WordPress and this sort of, you know, the decentralized or at least distributed web is that you don’t enjoy things like automatic backup. So if you forget to pay your hosting fees or lose your control panel access or you get hacked, etc. You can lose the site and all of your data, all of the content that you’ve worked hard to create. And sometimes you get lucky, you can pull bits from Google Crawl cache or archive. You know, there are some archiving tools on the web, but link crop remains a major problem. I mean, I looked at this for a piece of work I was doing internally and I was surprised. I mean even in legal like casework there’s a very high percentage of links that are dead, like links where the content no longer exists and these are citations in legal cases. I think it was something like 30%. So that’s the sort of thing that Arweave can help address by making sure that there is this kind of Internet native archiving built in. You can’t just lose a site and lose all of the content as a consequence.

Sam Williams:
Yeah, exactly. I think that study even was looking at Supreme Court cases. These are not like.

Dave Lockie:
Yes, that’s right.

Sam Williams:
I mean nothing in the legal. Yeah, nothing in the legal domain is trivial but that’s got to be the maximally untrivial venue. And yes, it’s interesting because I guess as humans we’re, we’re quite, we find it quite difficult to, to imagine like long term, long tail problems. You know, if something works today generally we’re, you know, we sort of get on with whatever we’ve got to do next. But when you look back at, you know, the Internet from 20 years ago, I think it’s something like 1.5 to 2% of those links still work and report to the same content as they did 20 years ago. It’s virtually nothing. But we don’t really think about this on a day to day basis because well, 20 years is a long time. Right. And yet. So there are ways that people can use. If you have content on your WordPress site today that is sort of external in some form, you can just replace those links with Arweave links and then at least that content is always around. Yeah. And that’s what we’ve been helping people do in various different forms now for nearly a decade and yeah, excited to see in 20 years time when all of those links still work, that’ll be. Yeah. Would you say a sense of accomplishment? Eight years is good but you know, we want much longer so.

Dave Lockie:
Yeah, definitely. So I’ve played around with Arweave since Vibe coding became possible and even over the last year or two I’ve seen a change in the way that we’ve worked. So let’s just touch on the, the way that one pays for Arweave so as like in a normal web, two like the usual web services that we pay for domains and Hosting and Amazon S3, you pay on an ongoing basis to keep those services active. So every month if I don’t pay my AWS bill then they will eventually close my account, delete my data and that data is gone forever. Whereas with Arweave, it’s a pay once, store forever model. And that’s a really crucial difference because it means that your credit card expires or you lose your PayPal password or whatever it is, you don’t then lose the content that you’ve saved to the weave it’s called. Right, the underlying data structure.

Sam Williams:
Yeah, that’s right, yeah. So. Well, personally, every time I hear the word forever, I cringe a little bit because we’re very, very serious about the mission here, which there’s no elegant, precise definition for in English or a single word that relates to it. But the goal is to store data for the maximum possible time period. And we really mean maximum. Yeah, forever. Inevitably, you know, Arweave will, before the heat death of the universe, it will no longer function as it does today. But it is intended to function for an extremely, extremely long period of time. And you’re absolutely right. The way that that works is through an endowment contribution when you upload your data rather than a single payment. So when you pay AWS or any other type of web host, you pay on a month to month basis, you stop paying and the service goes away. Whereas with Arweave, what you’re doing instead is in the same way that Bitcoin has a neutral monetary policy, so everyone can introspect it and essentially see when the tokens will be created and what the general sort of financial economy of the system is. You can do the same with Arweave, and we use that same property of trustless execution of a monetary policy, but we apply it to a different problem, which is, okay, you want to store data for the maximum possible time period. How do you do it? Well, you pay for 200 years worth of storage at present prices when you upload the data and put the tokens in in the first place, which sounds like it should be expensive, but actually the cost of storage is so low that yeah, 200 years is really, it works out at, you know, I think 12 to $15 per gig at the moment, something in that order. And so you don’t want to store like 4K security camera footage overnight somewhere you’re never going to look at. That’s not an appropriate use of the system. But for very valuable documents or writing text is extremely, extremely small. And so to store your blog post for 200 years plus worth of time at least then it costs like a fraction of a cent typically. So in certain or for certain types of data, it’s actually very, very affordable. And the interesting thing about the model is, okay, so you pay for 200 years worth of storage at present prices today. But as the cost of storage declines over time, the horizon, essentially the Runway of your data being stored in the system expands out. So if the cost of storage declines at a rate of 0.5% on average in a year, and you start with 200 years worth of storage power at the beginning of the year, by the end you’ve still got 200 years worth of storage power, storage purchasing power rather, which is very interesting effect. It’s essentially utilizing the deflationary effect of people’s ingenuity, I guess, in data encoding technologies to make the system sustainable for arbitrarily long time periods. And so it has a very, very different economic model. And the last thing to point out about that is probably that because Arweave buys its storage from a decentralized network of providers, anyone listening to this could go and boost up a node themselves now if they wanted to, and get paid for storing data. The barrier to entry is vastly, vastly different to starting a service like AWS. And because AWS, you have to gain trust from your users and you also have to gain distribution. Just people need to know about your service. And interestingly, yeah, Arweave buys its storage from its miners for about $0.78 per terabyte month, whereas AWS will sell you storage at a rate of about $23 per terabyte month on S3 for the equivalent tier of access rights. So while you’re paying a lot more for the upfront contribution rather than paying per month, actually, because they extort so so wildly in the centralized market, based really on that distribution and trust effect that they’ve managed to garner, it still ends up being not that much more expensive. I think it’s about 7 years or something. Cost on S3 equate to the cost of storage on RW permanently. And no one can tell you precisely how long that is, but the model is open and transparent. You can look at it yourself, you can read the papers and assess the risk and how long you think it will last. But yeah, it’s a radically different form of storage in that way.

Dave Lockie:
A lifetime is probably a reasonable way to think about it, right?

Sam Williams:
Ideally three lifetimes. That was the. I don’t think we succeeded unless it outlives 200 years. And interestingly, we picked that number because we felt that, well, people that are skeptical about this and I would understand that, I think had I been on the outside, I would also be, oh yeah, permanent storage. How’s that work? Yeah, I’d be suspicious. Well, if the price of storage never declines again, that 200 years is three lifetimes. It’s pretty damn permanent as it goes. But we actually expect that the price of storage will decline quite significantly in the future. Has a very, very long track record of doing so. Even before digital storage medium came about actually. But that’s a whole separate conversation that.

Dave Lockie:
Sounds like a rabbit hole for you, Mia. And a beer at some point.

Sam Williams:
Sure.

Dave Lockie:
So let’s switch gears slightly and talk about distribution. So one of the reasons AWS is such a massive profit center for Amazon is that they made a very well originally it used to be, you know, quick, easy and simple to spin up like infrastructure on demand if you needed compute or storage or whatever. You know, there’s APIs and everything is designed to be automatable and scalable and everything else. Now I referred back to my own kind of vibe coding stuff with Arweave and a WordPress plugin. It was remarkably simple to get working in that kind of simple concept of it, by the way. But over the last year I’ve noticed that there are now bundlers and if you upload to Arweave through one of these bundling services, you can upload up to 100 kilobytes for free. I mean, that’s how cheap it is to. I mean like by inference that’s how cheap it is to store on Arweave if you’re pushing something up like a text object, like a typical WordPress post. If you just take text and push it up, that’s going to be less than 100 kilobytes. So people could essentially back up their, their WordPress posts for free to Arweave right now.

Sam Williams:
Yes, that’s true. The network doesn’t subsidize that. But forward research, our company does at this point right now, in the future, you know, five years down the line, maybe it doesn’t quite work like that. But yeah, for now there’s. You’re right. 100 kilobytes is so cheap on Arweave we can happily subsidize it and doesn’t even touch our. Yeah, it’s very, very small part of.

Dave Lockie:
Our operating model and the developer tooling. I guess now AWS and other API first business, you know, Stripe and everyone else have kind of paved the way for how to build great developer experience. You’ve managed to do the same thing with Arweave with, you know, a heck of a lot lower budget than AWS needed to spin that all those concepts up from scratch. So I guess what I’m trying To say is a developer shouldn’t be put off by the idea that this is like new fancy crypto web three stuff in terms of actually using it. It’s all the tooling that you’re used to. You can mess about with it for free, certainly for now. And there are a number of GUI based tools, tools, you know, but just sort of not even needing any code that you can use to go and upload stuff to Arweave and mess about with it. So it’s very, very accessible right now, right?

Sam Williams:
Yeah. I actually think that building decentralized web applications should in its eventual form, and I think we’re pretty close already, depending on the complexity of the thing you want to build, should be far easier than building on the centralized web because infrastructure like it’s, it’s annoying to run and if it goes down then you got to go fix it. Whereas here we have open, mutual and sort of self organizing infrastructure that replicates the data around and makes sure your site or your whatever it is you’re uploading your content is available from many, many, many places. And so it really should be much easier to build on the decentralized web in its eventual form than it would be to deal with. I don’t know about you, but every time I look at the AWS management console, I sort of die a little bit inside. It’s unboundedly complex and I think that that probably stems from the fact they just haven’t had real competition. I mean, they’re so entrenched that they can afford for the product to be terrible and overpriced. But people are paying for trust and distribution, essentially.

Dave Lockie:
Which nobody gets fired for buying AWS.

Sam Williams:
Right, right, exactly, exactly. Except actually far more trustworthy is just open source software that you can audit yourself and understand how it works and you can read and assess the economic model for yourself. Like you are software in the system in a way that really you never will be with AWS and friends. And so yes, it’s very easy to get started with today. Like if you’re building a full web application and there’s a bunch of tools to help you deploy it to Arweave easily. And yeah, most of the time it’s completely free. And if you do end up paying, you know, go over 100 kilobytes and you do end up paying for it, you can also pay with credit cards and debit cards around the world. And basically what happens in the back end is that there’s a service, one would be Turbo from the ARIA team, that just converts Basically buys tokens and uploads the data for you. And you can pay them with Stripe. It’s very simple. Feels like using a Web2 service. And if you want to get started just uploading with a ui, you can use R Drive, for example. So if you just go to rDrive, well, actually any Arweave Gateway, you can go to rDrive.arweave.net or rDrive at IO or rDrive G8 way IO I think is another one. There’s about 400 of these things. It’s a properly decentralized web application, just lets you drag and drop and deploy any kinds of content that you might want to use in your site and encounter or embed in the same way as you normally would.

Dave Lockie:
Awesome. So if you’re building for WordPress right now, you can use Arwid for backups just using drag and drop. You know, if you need to do a migration, you’ve got a bunch of content somewhere, you’re exporting it, you’re putting it somewhere else. You, you can stick that back up on Arweave, it’s there as long as you need it and it’s very cheap to do. And you don’t need any coding skills, you don’t need any development, you just go to a website, upload it, and Bob’s your uncle, which is cool. Or if you want to get your hands dirty, or rather probably Claude or Cursa’s hands dirty, then you can build an integration so that for example, every time you post a WordPress post, it will push that content up into the weave as a backup. And you can imagine doing that for say WordPress has a built in versioning system, but that takes up space on disk. And if you have a lot of content and a lot of versions, that can start slowing down your WordPress database, if you’re offloading that versioning to Arweave, then you can stay very nimble in terms of your on site architecture. And I guess the same goes for attachment storage as well. If you don’t want to rely on your, I’ve got to say, WordPress.com the hosting account. But you know, your, wherever you host your WordPress, if you don’t want to clog that up with all of the media, the video, the images, the audio that you upload, then you could offload the media to Arweave and just embed the audio, video or picture like that. You know, just make it a media library integration so that you’re keeping your WordPress site itself, I guess entirely resilient. And to have these redundancies, you know, you’ve got your media offloaded, you’ve got your content backed up. And so if your site falls over, then you can just reconstruct it because all of the underlying elements are out there. And that’s assuming that you’re not just pushing like a whole backup up to Arweave on a daily or weekly basis.

Sam Williams:
Yeah, exactly right. And you could just do that. You could export the site, upload the, the archive that you get out of it, and then just load from that whenever you need to in the future as well.

Dave Lockie:
Now I’m going to switch gears a bit and move into the kind of like forward looking, I guess, beyond the basics and then kind of starting to look into the future. So one of the things that I love that Arweave have done is this concept of atomic assets. So let me see if I can explain that to our listeners and then you can tell me all the ways in which it was incomplete or incorrect. So most WordPress developers will be familiar with the idea of a WordPress post has a JSON object. You know, WordPress has a REST API built in. You can request a post, it gets returned as a JSON object and that object can then be used for whatever purpose. Like a typical purpose would be to have like a headless WordPress site where WordPress the CMS managing the content, but the content is then being accessed from and displayed on a totally separate like static front end, you know, all powered by JavaScript. So I guess conceptually, if you take that WordPress post, you structure it slightly differently and you add a few extra fields and push it up to Arweave. What you end up with is something called an atomic asset. And an atomic asset is a particular standard for storing stuff on Arweave. And its aim is to make that piece of content, that asset, complete in a totally Standalone basis. So what do I mean by that? When you publish a WordPress post, that post is tied to a specific URL that is, you know, part of a domain and then a site structure. It also typically doesn’t contain information about the author beyond who they are on that WordPress site. It doesn’t have a concept of like a global author. It typically doesn’t include usage rights, such as, is it Creative Commons? Is it copyright? How like, can other people use this? And if so, how? And it also often doesn’t include a concept of monetization, like, okay, you can use this content, but I would like to be paid on this basis and this is how you do it and how much it costs. I’m probably leaving out a bunch. But that’s the idea, right, is to like, bundle all the things that somebody interacting with a piece of content might know, might want to know into one package so that you can use it, you can remunerate the author, you can contact the author, you can see what else that content relates to, and you can display it in your own experience in a way that is both complete and respectful of all of the connections and metadata. I guess that that post has. Tell me how I can improve that mental model.

Sam Williams:
Yeah, no, you can’t see me, unfortunately, but I am nodding. This was a great explanation. Yeah. The basic idea is that content on the Internet, it’s not just bytes, it’s bytes plus a bunch of metadata that really is necessary in order to understand what the content is. And you’re absolutely right to put the emphasis on authorship. Right. You know, the. The content that’s available is openly, would you say, addressable, which is great, but you have to be respectful of, well, who made it and you know, what, what ways would they like to make be remunerated, if at all, for the content that they’ve created. And so an atomic asset basically lets you upload a piece of content, have it be displayable in any other application, but it moves atomically, hence the name, with all of the information that is necessary to display it correctly. And that includes who to pay for access to the content. It’s quite interesting if you have an atomic asset, if you send someone, the link inherent within the link is the. Yeah, the sort of ownership rights of the content itself. So anyone can access this content, for example, might be one rule. But you should please pay, you know, X micro dollars, I suppose, in order to do so. And what we’re building towards with this is a web where the sort of platform creator relationship is turned inside out to the centralized web. So on the centralized web, I mean, the audience here, publishing with WordPress, have evidently seen this problem and are working around it already to some extent. But if you were to publish on x or on YouTube or any of these other large platforms, you give the right for YouTube and whoever that platform is to distribute your content and then basically give you some amount of money back that they get to determine and they get to change the rules around that at any time. And that gives the platforms an enormous amount of power. Because if you build up an audience on those platforms, well, your access to that audience is completely dependent on that company and by the nature of companies themselves. And I don’t see any moral failing with this, it’s just the nature of the sort of deal, the agreement that has been made between the parties. Well, that company is there to extract as much value as it can from its audience. And you know, then there’s a market where they’re basically providing a service that is useful to people, but they’re also trying to extract as much profit as they can from it. This leads, I mean this would you say, the elegance of that market structure works only questionably, really, when there are such large monopolies, because practically speaking, you have quite a difficult time building up an audience outside of them. And so instead what we see as well, content is the bedrock of the web. What if we made it so that the platforms were consumers of the content? They queried an open database, which arbee is you can query it and ask for, hey, I’d like all of the content that is, I know content type, JPEG or something and has some tags related to dogs or whatever it is you happen to want. And then you can display it on your site, but your site is just a sort of pass through basically for the connection between the creator and the audience, the person actually accessing the content. And as a developer, this is really exciting too because you don’t have the cold start problem, right? Every time you start a new application, you’re starting with an empty database which you have to then fashion into something that is worthy of the attention of the audience. Here it’s the other way around. You start with all of the data on the web or the perma web, we call it the web inside Arweave. It’s about 20 billion pieces of information that spans a huge number of different topics and say types of content and all the rest. You start with that as your database on day one. And then you can make an experience that’s much more exciting and engaging. For people just building on that data that’s already available. And as a content creator, you can upload to this. So this network is not a platform. There’s no company that intermediates it. We, our company started the network, but we certainly don’t run it today. It’s neutrally held public infrastructure basically at this point. And you can set the access rights. So then when a platform comes along, they do want to use your content. Well, sure, they can do, but there’s a legal license associated with it that that tells them the ways that they can do that and how they can pay you for the right to do so. So yeah, we’re really trying to move towards this model of the web where it’s content native and then applications and just thin layers on top. And to your point about Vibe coding, I think this is, would you say the time is now for this type of model? The world where like you build a little web application, then it becomes a company because you have to host all the infrastructure and you run it like a business that just really isn’t necessary, necessary, necessary necessarily anymore because you know, you can throw together something fun and cool in Claude, you build on a huge database of existing content and it’s a fun experience that people can use. You Peter uploaded once the web UI is then essentially available in many, many different places your users can find it over time. You as a person uploaded it, don’t have any usage fees related to the application. Yeah, I think we’re moving to a world where the application side of things, the user interfaces become thinner and thinner and less relevant and really it’s all about the content. And that’s what Arrive is giant open content bank that yeah. Allows you to be paid for, for making great content that other people want to access.

Dave Lockie:
Yeah, I think the cold start solution is really interesting. There’s a couple of other things that I’ve been thinking about more recently as well. One is around the importance of relationships and of nodes versus edges. And you could think about that as, you know, a content is a represent. It kind of represents an edge. So if you’ve got two people in a network, then often it’s a piece of content that connects them rather than there being like, I guess you have a direct relationship with some people, but typically online, that is through a piece of content that somebody creates and somebody reads or somebody collects or buys or whatever it is. So there’s also this inherent social network and I use that term kind of advisedly, but there’s also this inherent like people network in Arweave. And there’s also an interesting dynamic in terms of content liability as well. Now, the big social media platforms have issues with content moderation, certainly when they can’t buy their way out of those issues. But because the content that you’re building on in Arweave isn’t necessarily hosted. Like you’re not hosting it, right? You’re not uploading it, you’re not hosting it. That’s the direct responsibility of the users. I think there’s. And obviously that, you know, law stills apply and there’s got to. You can’t just kind of ignore that stuff, but it changes the, I guess, like the cost and overhead of dealing with that content stuff as well. And it does not say, are we visa Wild West? You have content takedown processes and like that is both the network and the individual author’s responsibility as a builder. Maybe not so much in the way that it would be if you were say, building that t network so that women could share intelligence about men. Like you wouldn’t necessarily have the same kind of data liability that you would if you were building on Arweave. Is that fair?

Sam Williams:
Yes, exactly. So the liability is with the data stores, the nodes in the system, and the way that the process works is that any node is allowed to introspect the data. The data is by default unencrypted. You can choose to encrypt it if you want, but you are liable for the data that you store as a node in Arweave. And then to navigate this. Yeah, miners. So miners are the people that store data in Arweave. Kind of like Bitcoin miners, except they’re doing something useful with their, with their expenditure rather than just burning energy. Yes. The node operators, they are the people that are responsible for the content that they store. The platform, which really is actually just data. Right. It’s HTML, CSS, JavaScript bundled together. It’s not upon them to moderate things because they’re not storing or hosting any data. In fact, arguably they’re not even helping people upload it. It’s just a user interface on top of services that other people run. And yes, those miners, they get together and they build these content policy lists so you can send a message to the network if you want, that says please don’t store this piece of content. I know, for a GDPR request or similar. And then in localities where GDPR applies, the miners will make sure not to replicate that information because it is against the law to do so.

Dave Lockie:
Their legal responsibility.

Sam Williams:
Exactly, exactly. Yeah. And they self organize into these groups that have different content policies. And it works well. It’s also intended to. The core reason we started Arweave in the first place was because we wanted there to be a system that was resilient to political censorship on the Internet. We saw, and unfortunately are seeing, that the Internet in the west is becoming more and more closed, like some countries elsewhere in the world. And we wanted to make sure that there was a mechanism where, yeah, politically important content, the types of things that people would want to make sure, become part of the historical record for future humans to look back on and to understand how our world today worked and what people were thinking and feeling at that time. They can store it in a network that will distribute it across so many geographies that practically speaking, political speech would never really be censored because that which is legal and say, Turkey, just give one random example, it’s not illegal in China and vice versa. However, there is certain types of content that virtually universally is deemed to be reprehensible and subsequently no one in the network will store it. And so it has a nice sort of mechanism for dealing with this complex problem of what content should become part of the permanent record of humanity. And yeah, and it doesn’t require a centralized controller. Our company, if we said tomorrow, which we definitely wouldn’t do, but you know, no Xi Jinping Winnie the Pooh pictures, it wouldn’t matter. We barely run any nodes at all. In fact, we just run them for testing purposes. It would be up to the node operators to all decide, well, Xi Jinping, Winnie the Pooh is not something we want to store, which is very, very hard to imagine would really happen. So you have trustlessness of data replication over time as a consequence of the decentralized nature of decision making. Basically you would need everyone to make a terrible error at the same moment in order for the content not to be stored, which for almost all types of content really is not an applicable problem. And so, yeah, that was the core thing we were setting out to solve.

Dave Lockie:
Yeah, amazing. There’s a couple of things on my mind. I also want to make sure I leave you some time to get over anything that you think the WordPress community needs to hear. I want to share first a mentor model, which is at the moment we think of content as belonging to a website, of being subservient to a website. You go to the website and then on to the content. You know, like the domain comes first and then the content is like the path off the domain. And for me there’s like this I guess the original sin or the root cause of a lot of the centralization of power that we’ve seen online started with databases. Like it’s. There is until arguably until Bitcoin, there wasn’t really a concept of an open source database like an open database. And of course you can run, you can open source code, you can download it and they run it. But as soon as you download it and you need to connect it to a database and store data, then like suddenly the amount of liability and responsibility and like the difficulty of doing that in an open web way just increases exponentially. And so that’s when it then became attractive for people to use services that could provide databases. Because it’s another technology layer. It’s kind of difficult. They need to be fast, they need to be. The whole nature of them draws towards this sort of centralized paradigm. And one of the things that really excites me about Arweave is the idea to kind of turn that on its head so that the content lives in what’s essentially like an open database. Like, like open source code. Anyone can copy and use and do stuff with. An open database is there for anyone to query and retrieve the data and push data to. And it’s not lawless. As we’ve talked about. There are both legal and economic constraints and incentives to manage that behavior. But you kind of flip it on its head and you think about, well, maybe the, maybe the role of a website in the future is there. So yes, it’s there to help people build and create the content, but perhaps they don’t store it locally. Perhaps what they do is they push it into the open database and then what happens on their website is a curation of that content. You know, content they’ve created on that site, but also potentially content they’ve created on a different platform. Could be open source, like Ghost, could be a centralized platform like X or Facebook, but somebody’s taken that post and saved it to Arweave. So you end up with the relationship between the content and the website sort of flipping, but also taking on this kind of richer and more complex relationship. So, yeah, I don’t know if you have any comment on that. And then I wanted to talk about AO as well, just to give you like an idea of time and everything else.

Sam Williams:
Yeah, of course. No, that was perfect. The only thing I would possibly think to add to it is that the way we see it, the web initially envisaged, and certainly the web that I grew up with, I’m 32 now, so I grew up around so young.

Dave Lockie:
It’s ridiculous.

Sam Williams:
Well, I. The web I remember when I was younger was much wider and more varied and I. Yeah, that’s why the WordPress community, I feel, is so kind of aligned with what we’re working on here, which is really maintaining an open platform or an open network of content that people can remix and use in different ways that seem to be inherent to the original vision of the web and has crumbled to the point of. Yeah, I really think WordPress is about the only bastion left that is extremely widely used. And if we don’t solve this problem now, I think we are heading towards just tighter and tighter controls of the types of experiences people can have on the web. Like really the browser, about five to maybe 10 years ago even became just like the dominant application distribution platform. You know, HTML, CSS, JavaScript bundled together is just the way that you write what are more or less desktop applications. And then the number of desktop applications became extremely tight and limited. I don’t know about you, but yeah, I feel passionately that if we could have a world where the content was dissociated from the application, then the application space could be much, much wider and iterate in a far, far faster rate and then we would get back some of that spirit that existed in the early web and hopefully something, something new as well. So, yeah, I just wholeheartedly agree as first.

Dave Lockie:
Great. Yeah, it’s daunting to think about how to take on the ludicrous amount of power that the social platforms have. And like, I don’t mean to vilify them, you know, you know, I would be hypocritical to do so. I definitely use them, but that it just makes me more passionate about trying to find ways for the open web to stay competitive in a way that is true to its values and the humane mission and vision of the Internet and open source technology. So we’ve talked about something that is quite complex in concept, actually very accessible in terms of utility and has this tremendous value add now and also promise for more going forward. But you didn’t stop there, Sam, with all of your. You said 32 years, right? You didn’t stop building like a global permanent information storage protocol. You then layered on top of it something called so. Tell us about ao, Sam.

Sam Williams:
Yeah, so sort of continuation of the same mission. And it uses Arweave as the base layer. Essentially what we saw is that, okay, Arweave is an open data lake. You can query it in the using GraphQL. Actually there are other querying layers as well. But that’s the main one that people use. And so it feels like a database to use. And you can store your web application front end inside it and it looks and feels like you’re using a web application. It’s just kind of magic that you can also go to. I think it’s about 400 gateways at the moment. Yeah, 400 different domain names and have the same experience surf back to you. The limitation with Arweave before AO was okay, well, what if I want to do like lots of complex processing, right? An amount of computation that can’t be done on the client side and maybe actually it’s just kind of like a bit of a nuisance to do on the client side. It’s theoretically possible, but people want easy developer experience, understandably. What if I have real compute requirements for my web application? Well, there wasn’t really a technology in the decentralized web space that allowed you to do that. And so that’s essentially what we built with ao. It’s probably bit too complex to get into in detail now, but broadly speaking, it is a decentralized computation network that allows you to run your compute workloads on just like are we even open set of nodes that you don’t have to have trust in. So you can replace the centralized parts of any service with decentralized computation, which is available at the prices that people are willing to forego them at which as we discussed as relates to storage, is far, far, far cheaper than what you would pay Amazon and friends and have that workload executed in a way that you can trust reliably the result. And so this really allows you to go from, you know, sort of create, read, update, delete style web applications to full stack, you know, whatever you can imagine you can build in a decentralized way. And we see it as just like the next inevitable step on this process of building a stack that can replace the centralized technologies that have become dominant on the Internet today.

Dave Lockie:
Amazing. And I think that’s probably enough to melt most people’s brains for now. So perhaps if we can find a slot, you could come on and give us a bit of a deeper dive into AO. Hopefully today’s given WordPress developers a idea about how and why to use Arweave and that it’s like totally easy and cheap to do that. Is there anything else that you’d like people to walk away from today?

Sam Williams:
No, I mean, that was great. It was a very interesting conversation. Yeah. And lots to cover in the future perhaps as well. I just an impassioned if not now, when? I would say we ought to ride this ship. Therefore there are no people like us left that even remember the web of the early days.

Dave Lockie:
If not now, then when? I agree with you, Sam, and I’m sure lots of our listeners do as well. I know how busy you are. I really appreciate. Well, I really admire everything that you’ve managed to build, and I really appreciate you giving up some of your time to come and help the WordPress WooCommerce open web community understand better. And I hope it sends a whole bunch of more people creating atomic assets so that we can preserve the vitality of the open web.

Sam Williams:
Amazing. Thank you very much.

Dave Lockie:
Thanks Sam.