Reclaiming Digital Public Squares in the Botnet Era | Taylor Hulsmans - Hashchan

All right. Well, thank you for F Denver for giving me the speaker slot and thanks everyone for finding themselves at this presentation. Certainly I'm excited to give it. Uh my name's Taylor. I go by Polar Saticus online.

And I really enjoy building and thinking about decentralized social media, how we can create systems that align to the core objectives or at least the core objectives when the internet was created, uh which ties into the topic of the presentation I want to give today called digital public spaces in the age of the AI botnet swarm and how can we reclaim them? Uh for brevity, I define a digital public space as an area where you don't need to log in in order to view content. You don't need to hand over anything to be able to write to that database or any of that sort. So, it's very much, you know, like a public square uh inside of a kind of a private enterprise. So, I thought I'd divide this talk into two sections.

Uh maybe spend the first four or five minutes talking about the history of why we need a re-architecture just to enjoy the basic services that we've become accustomed to online. And then through a case study called Hashan, we'll go into these architectural decisions and how you guys can apply them in your apps down the line. So, to begin, there was a dream that was the internet. If Tim Berners-Lee even whispered it at CERN, it would simply poof and it vanish. It spoke to a universal access to the world's information.

There was a time where we typed in www. to go to a URL. I don't know. I haven't done that in a long time, though. We could communicate in real time with each other with complex data types.

Perhaps it's the metaphysical postulate of the information age that we affect the world by what we choose to browse. That's Tim Berners-Lee. And I'd like to ask you guys, what does it mean if only we can really browse bot spam? I'd like to pinpoint why this the beginning of why this architecture Oh, can we get a Is that coming through? Okay, I'll just talk then.

Yeah. Um I'd like to begin this talk with the fallout of monetization. It really began in earnest with web 2 monetizing the net. Advertisers are going to comment sections instead of buying ad space because it was just simply more effective. We saw advertisers in asking these platforms to say, "Hey, can we just have more user data so we can go better target these ads?"

We saw click farmers really get into the mix on everything and say, "Hey, you know, we're going to farm engagement here, get the advertiser revenue. We're going to give fake likes to other people." And so, this really created this self-perpetuating cycle that I'd say around 2016 was really picked up by political campaigns that began using the techniques and the information to create high-dimensional neural nets for the purposes of behaviorally nudging people into team A or team B. Needless to say, the cat was out of the bag. And we saw a paranational struggle for narrative dominance really ensue.

Uh so, I think it's really natural that the users at the end of the line were the most affected. Uh but in second place, it was the moderators and the sysadmins of those platforms that got subjected to like employee infiltration, bribery of the sysadmins, corrupted moderation teams. So, there's no shortage of people trying to create solutions to this. Um but for our purposes in the open internet, they're really Faustian. You would get one thing, but it would cost double the other.

A good example being Cloudflare is that they provide you DDoS detection from these botnet swarms, but it would cost the users the user data. You get the Cloudflare cookies, game over. On the other side, you try to provide free a space free from data mining, uh but then you're just subjected to flooding and content and spam. Um we saw advertisers become particularly choosy about which content they were associated with, even these comment sections. Uh so, we saw this departure of free speech being broken into this moniker, free speech is not reach.

Uh and we saw the beginning of the shadow and the heaven banning and the beginning of the splinternet. Uh nowadays, you need to fork over your your phone number to use close AI. You need to have a Google account place it over if you even want to view a unmoderated section of Reddit. Um search engines as well. I don't know if you guys agree, but it's a shadow of what it was a decade ago.

And this could be pertained to many reasons, but SEO option optimization and gaming the algorithm are just a couple of them. So, needless to say, if the internet isn't dead already, then it's certainly dying. Uh a homage to the dead internet theory that has made its rounds a couple years ago. Uh with that explores this idea that more and more so by each passing day, we're interacting with bots instead of having a human-to-human interaction. And I don't know, too.

Like, can you guys feel it? Is like there's this artificial inducement of memes, this repeating of content, this tyranny of the present in the some sort of essence where you can't really find anything that hasn't existed more than three or four months ago. And if one thing is forgotten about, we're just onto the next thing and the next thing. Like, I swear they like purloined the most like darkest, deepest corners of psychology to find these most hypnotic-inducing algorithms. So, if there is a live internet maybe existing on the fringes, uh it's really hard physiologically to get there through the doom scroll.

So, I'd like to call this term the bot swarm gulag, where we show up for a service, but as each year goes on, we're just provided more of an illusion of that service, where the primary goal isn't talking and communicating with our communities, our nations, our friends, our family, but it's basically to provide data for the companies or get force-fed content by those companies. Uh so, we're there really in essence to train their botnets or be trained by their botnets. Um so, the image of a digital dairy farm where needless to say, we're not the farmers, we're the cows. So, that kind of ends it with the history piece. I want to go into architectural decisions now that we can take that ideally brings life back to the open digital places of the internet.

And I chose open access forums cuz they are really uniquely situated in a space where they're helpless. Uh we hear a lot, you know, especially in the web 3 days uh these days about decentralized identity, DIDs. You know, provide a retina scan to uh something, provide a palm scan to something else. Um in open access, we're just but we don't have a digital scanner of anything. You can't even use those.

They're not even Faustian. They're just non-solutions at all. Uh so, the using tools kit doesn't really work. And also maybe as well, like there's a hope at least a couple years ago that the bot spam that we were receiving might actually improve these spaces. Like we'd have these philosopher AIs going around to all the corners of the net providing great context, providing great reasoning.

Uh but we've seen from different I know, Crest Research, I think was one that put on is that we're just going to flood these spaces with drivel with the hopes of just guiding these individuals into more monetizable portions of the net. So, I think uh one of my favorite quotes, I don't know if it was Descartes or Voltaire, but they say uh if any community gets its laughs by pretending to be idiots, they'll eventually be flooded by other other idiots who mistakenly believe they're in good company. And so, yeah. If we're mimicking the idealisms that are being dominated and there's a slight group of individuals deploying bot swarms dominating the signal-to-noise ratio, like we got to be really careful about imitating and leading up with a lot of drivel we see in these spaces. So, to begin, I'd like to take you through the journey of a user to a particular website starting from the very beginning.

When a user logs on, they'll always go to a DNS record behind a static IP. We can get rid of this. We can just simply give a website in a zip file and have them interact with distributed databases either on a high end with their own RPC provider or, you know, actually like go out to all the people building homes and say, "Hey, you know, you have a furnace. Like, let's give them like 20 terabytes and their own node." And so, just give them an archive node.

And so, Nick Szabo said it best, uh trusted third parties are security holes. By eliminating the need for a DNS record and a static IP, like we get rid of number one, the IP logging, the MAC address hard logging, the Cloudflare cookies, the Google cookies, whatever's happening behind the scenes on these sites. We talk about modern infiltration, sysadmin infiltration, creating like subgraphs of individuals as well. Like we see one graph on the surface, whatever its graph is being deployed underneath the surface remains a question. We can get rid of all that.

And because invalid transactions aren't propagated by the mempool, we also receive a degree of horizontal scaling against DDoS. This means we can get rid of Cloudflare as well. And also on a nice flip side, as we saw as with Internet Archive, now that all messages are being propagated at block speed, we get archiving comes free. And because everything is cryptographically signed, we get anti-shadow anti-shadow banning guarantees. Uh yes.

The second point is that normally when a user hits this website, they'll be asked to log in. Uh lots of stuff being asked here about creating portable identities. I think we can do one greater. We can create togglable identities. We can allow the user object to become amorphous in essence and choosable by that.

You say portable identity requires these five things. We can say, "Hey, you don't need them at all. You can bring one. You can bring five." And so just we can complete we can accomplish this just by how they choose to fund their wallet address.

And so at the high end if you wanted to create like a Reddit karma service, you could say, "Hey, you know, you have to have an ENS record. You need to have a a year's worth of posting history." If you want to go a level below that, you can say, "Hey, you know, just fund a fresh wallet through a KYC exchange." Probably roughly equivalent to the centralized services we have today. Whole slew down there around non-KYC C exchanges all down to the tail end for this at least about using a chip mixer for a fresh wallet.

So, one of the probably core issues around what I was curtailing with is how do you get rid of the bot spam? Uh, without these centralized services. And what I found is that simply by having a transaction fee associated with every single post, we get a natural tuning friction. And I and I'll do as an addendum here as well as that who wants to pay to post? I don't think anybody.

Uh, but I think and this is the better hash chain as well is that it will be this sort of golden age that we tell people in 10 years time that is like, "Oh yeah, we used to have clear comms throughout the net." And like that will just be like, "Whoa, really?" Like we never enjoyed that before. It's like so in order to have clear comms in the next decade, something will have to happen. I think this is the best way.

But we can see that just by providing a small transaction fee that an individual will pay post pay to post some one thing, but a floater who wants to engage in any sort of strategy around spamming or shilling or flooding saying, "Hey, I need to post a hundred times, a thousand times just to execute my strategy." They'll see that order of magnitude cost increase go. In the web two world today, as long as you as you got your botnet set up, the nice set of IPs distributed, the cost of flooding is the same as cost of posting. And so you can go, "Okay, you can say, 'Hey, I'm on a really cheap chain. I've seen like 10 to the negative cents per post on these lighter chains.

So like doing it through the setup we'll talk soon. It's like really really cheap. But if you're still getting a lot of spam, you can go to a heavier chain such as Ethereum.'" And then you can see that order of magnitude scale up really amplify. So like if it's on a small chain, it's like maybe 1,500.

If it's on Ethereum and it's a dollar to post, you get 15,000. And so it's like huge difference. And like I think we saw it a week ago as well that if an individual really wants to tell the truth on something, they can start burning their ether. So if you want to like burn 1.3 million of ether to say something, this is the way to do it.

This way it's easy. So we talk about writing to the database. Now that Axis talking about reading to the database. And we talked had a really good conversation about from AWS about implementing the graph nodes. Um, I want to do the dirty solution.

And that's client-side event log indexing. We don't need the catch-all generic solution if you can fit your use case to a client-side index. So for example, for a social network, if they want to look at a thread, they only have to index the thread. They don't have to index the entire blockchain. Like what is what's the cost of a chain reorg?

Likely minimal. And so this can come down to the user choosing an RPC that suits their needs, getting their own archive node. Uh, I tried this a couple years ago and I'll use to say it's a lot better place today than it is three years ago. Infura did a wonderful job indexing the event logs themselves. I haven't even come across a block range index for Ethereum with them.

Quite a different story from other chains, but by using event log and client-side indexing, we can just pull the data as needed to one person's use case and go from there. Uh, on the flip side as well, incentive structures can exist that like just say, "Okay, I indexed it. I'm going to float it into the atmosphere and someone else that shows up on the net can pull that, do a quick verify, or just trust it depending on your needs." And so that leads so we have some text committed. We got to talk about the images, right?

Too heavy to place on chain. So we use IPFS here, but the main point that I want to talk about is about handling individual and platform liability. Uh, the unique system that we have here at Hash chain is that it makes it non-platform and non-platform or non-user liability. Uh, we've seen in these edges of the net that illegal content, DCMA, and copyright are really really potent lawfare techniques uh, to take a site down. Especially for the smaller ones, they have a really limited budget, really limited time.

All you got to do is really just find out and say, "Hey, this person has 10 hours to moderate. Let's just do 11 hours worth of content damage." So by making an individual responsible for their own content, we can bring that liability off the platform. Uh, on the peer-to-peer side it's also been an attack vector. Uh, we saw perhaps like from ZeroNet back in those days that if you become responsible for hosting the image content of another user on the node, then you're just starting to propagate these images.

You don't know where they're from and it is a huge liability. Uh, so this solves that as well from both the other peer side. Uh, users can then opt in to pinning other users' images to create liveliness on the network. And to also if you want to create an incentive structure around this that increases liveliness, it's a natural way to go forward. Um, so now that we have this kind of idea around images and how we're protecting ourselves from different types of illegal images, we talk about the moderation services.

Um, Blue Sky did a really good job with their pluggable moderation services even though they do I believe to best of my knowledge hardcode their moderation service into their client. Uh, and there's some other aspects of that tech stack that are super expensive uh, to implement for another user that just wants to show up. Uh, in systems like this is pretty much you can go. Uh, but we want to make pluggable moderation services to handle the moral hazard of corrupt sysadmins and overzealous mods. So the idea is that also we don't want them to delete content.

We want them to provide filtering of content. And so in this way we provide and we maintain a unified graph of all transactions. So a user based on their own preferences can choose different moderation services. We can have different rating systems attached to those services and we provide a cryptographic link from a moderation services contract to the actual moderator. Uh, so the question will be how those moderations are actually being implemented and we can have the moderator do signed messages over WebRTC.

So they're looking at posts that are coming in and say, "We don't like this. We're going to sign it. We're going to use OrbitDB, provide authentication on top, and anyone who's subscribed to our moderation service will know cryptographically that the moderator is the moderator of that particular post." And so we can do this by a web socket, but it has a central point of failure. Uh, services like libp2p and OrbitDB do have WebRTC coming down the line and that means you only need a cluster or growing cluster of really really micro stun and turn services just to facilitate the initial connection.

Doesn't even have to be connected uh, to the original thing. At least to best of my knowledge. Um, so perhaps the last point that I want to touch on um, is that by associating a piece of content with an Ethereum address, it becomes fair use ready. And my fair use ready, we see the AI companies today just scoping the net out of of everything, doing the lawfare calculus, deny to fend a post, whatever. We're going to make a hundred billion on this.

What who cares if we pay 10 billion in legal costs. Uh, but the point being is that if fair use ever does come down the line, it's a lot easier for a company to say, "Hey, just push this 10 ether or whatever to this address." instead of trying to like facilitate a hundred or a thousand different Visa connections or whatever they need to do. Uh, so doing this model makes it a lot more suitable for retro PGF uh, and fair use attributions. Um, yeah.

So in conclusion, uh, the fallout from web two monetization has birthed this idea called the bot swarm gulag. Uh, here we're provided an illusion of these services where each by day they're actually getting more degraded uh, and we're just being sent there to train their bot swarms. The existing solution space that protects users from this sort of digital gulag effect is really really ill-suited to digital public spaces cuz just by definition we can't have a user account. So by using web three technologies, we can preserve digital public spaces and hopefully and ideally get back to the use of those those original core ideas that drove the original founders of the internet to even create it. So thank you.