DAPPCON 2025: Fileverse: Self-Custodial x Non-Financial x Usable - Andreas Tsamados

Transcript

Hello. Hello. Uh yeah, indeed. Whoa, that's a strong voice. Hello. Thanks for being here while uh the prophet is speaking as well. Um so there's a lot of things that I'm going to talk about today. I'm going to do it uh quite uh succinctly. I only have maybe four slides that are a little bit heavy on the on the words. Uh so it will be the first few slides. Uh don't worry after that it's mainly memes and good vibes but with of course important ideas being communicated.

Um so today I'm going to talk to you about uh subversive apps and safe communities. Okay, here are the slides and let's start with what I mean by that. So you'll see the next three slides is something that is connected with what I do on the side uh which is I'm finishing my PhD at the Oxford Internet Institute and over there we have a lab called the subversive technologies uh uh labs and we look at global information controls and uh uh censorship on the internet etc. So uh this is where uh let's say the conceptual foundation of what I'm going to talk about today comes from. So first

things first uh we're going to be exploring technologies that provide control over information flows and action on the internet and those that resist or subvert that control. This is a quote by uh just right uh from the University of Oxford. Uh I highly recommend checking him out. So um just to establish quickly the premise of uh my talk uh the internet has created a new landscape. I think we're all pretty aware uh for the communication and collaborative creation of information and has caused uh broad centralization uh of the global communications infrastructure. So as a result of that the technologies comprising the network

have come to represent a significant means of control over the creation, access and dissemination of large volumes of information and uh information and the means by which it is communicated and controlled are now really key concerns of uh uh of uh our times. And unfortunately although we have seen uh a proper uh uh a proper movement towards at least having permissionless money, we haven't seen it as significantly with data. And that's a shame. And my argument in this talk will be that we have exactly the

infrastructure and technology we need to actually make that move also for data. And actually that move makes complete sense to uh uh to enterprise companies as well. Uh and it's not just okay valuedriven you should create you should try this new technology because uh it has a better soul. No, even as a business there are some really concrete factors that you need to consider. I'll go over them a little bit later. So last uh last or two last slides with few too many words. Uh computer networks and the technologies um on which they are built have evolved be beyond their original

scope and purpose. It's quite important to understand that if you hear uh old architects of the internet uh if you listen to Bruce Nyer they say the same thing. Hey we never had privacy in in mind when we were creating those networks. And so now we're seeing a lot of those issues bite us in the ass. Um and two of those uh issues and this is why I talk about enterprise clients are privacy and information security. uh those things were not considered in the original design of core protocols and this has resulted in ongoing struggle to ensure that fundamentally flawed infrastructure can meet the requirements placed on it and its role as a global

information network. Okay, last slide with a chunk. Uh so through both its flaws and intended functionality, the internet has produced that ecosystem of cultures, institutions and technologies that either exert or resist and circumvent uh control over information. Now anybody knows what this is? It's the internet. It's the internet is the the TLDDR.

Those are uh um the internet cables that are powering all your connections. And looking like that, it looks like it's pretty decentralized. Uh but unfortunately, in the past decade, we've seen a lot of private companies, Google, Microsoft, etc., usually the same uh uh corporates uh centralizing even that uh fundamental core infrastructure of the internet. And it's a super worrying trend. Uh usually it used to go through consortia through uh non-governmental organizations collaborating with uh consortium of uh governments and a few private companies to lay down those very

expensive internet cables fiber optic cables which just as a data point it cost just to connect the north of Africa with the south of Europe. It's cables that are worth uh around $200 billion. So you know if you make one mistake that's it your company is bankrupt. So those those projects are quite important. Uh I'm also using that slide to uh to shill those two uh projects over here netblocks.org and uni.org. They're amazing applications for you to monitor global information controls censorship. Uh right now with the wars that I wars that are going around the world you can see very interesting. It's

something that's not talked about enough I think in in the media. You can see how internet is being shut down as a means of uh uh controlling a population making sure that information doesn't come out about the atrocities happening in a region etc. So track those things and you can contribute especially uni every time you do a scan of a local network it contributes openly to those communities that are tracking uh this information. So I started very low the uh submarine cables. Uh one layer up we see other trends of centralization uh on the internet. Uh it's important to remind ourselves that most of the internet is

really being accessed through mobile more and more as well. uh especially for the global south they are mainly mobile internet natives and it is quite important especially in web 3 to understand that trend otherwise we'll keep optimizing for the same you know ecochamber that is not really at least not for me the the core purpose of what we're doing in the space uh another important centralization trend is that there are really two operating softwares uh that are dominating that mobile space and it's Android and iOS.

Now, further up, uh this is pretty latest data from Cloudflare. Uh where is most of internet traffic going? Well, uh the usual culprits again. It's a Google, Facebook, it's Tik Tok, uh it's Microsoft, YouTube, and those things have been really quite established for the past 10 years. uh we see maybe a few apps come up and try to challenge their dominance but they they challenge that dominance with the exact same purposes which is to choke the market and create monopolies. So perfect.

Another very important data point is that um 55% of all the world's data is stored on the cloud. Uh I'll explain why I consider that catastrophic. And also 96% of all apps are cloud-based. This is usually the the slide that I like starting my uh my talks with. If you were at DAK or or DAPCON, um this is where I spend most of my time uh getting people's attention on which is quite dramatic in my perspective. It's that uh over half of all internet uh all of the globe's internet users trust one single closed source centralized app suite for

their daily knowledge uh creation and collaboration. It's catastrophic and the most important point I'm making today is that it's really not needed anymore. uh maybe 15 years ago where people wanted to move from uh Microsoft to something collaborative. This explained a lot of the move into cloud computing but those same arguments are obsolete and maybe worse than obsolete. We're kind of being dogmatic with our approach on uh with those technologies and using those technologies. It's uh I like using that image to uh to make a parallel here which is despite having constantly been on the

receiving end of the negative uh consequences of trusting centralized uh uh cloud empires. We still continue using them as if okay next time it's going to improve, next time it's going to get better. So it is pretty dogmatic. Sorry, I get I got a little bit overenthusiastic with the priests and monks blessing our machine overlords. So, you have extra ones here uh um for fun. Now, the problem speaks for itself. I I've said it in three different ways already, but just some concrete examples uh which is always uh important to show.

Last year we lost uh uh billions of dollars in the global economy because Microsoft Outlook was unavailable for 6 hours. Uh same thing uh uh um employee of Google decided to arbitrarily uh uh delete a bunch of data from a big Australian pension fund. They lost all their data, the backup as well and that's it. And the ridiculous part is that it was to be expected for a company that is worth that much. You start trying to dissociate yourself from those dependencies on third parties. And yet when we talk about, you know, data storage, communication, it seems that

we're not having the same reflexes we have for money and other uh many other problems killed by Google. There are over 460 projects that were killed by uh Google. There is a nice uh website to follow that uh blah blah blah. Something that I always do before I talk, I talked about it at DAG. So my my latest figures are from two weeks ago. If you go on some raid forum clones uh on the dark web, you'll see that uh you can buy 10,000 people's uh personal information that were leaked on a regular basis from those uh uh big cloud empires and you can buy 10,000 uh accounts personal information for approximately $6.

So um you'll see that a lot of companies a lot of organizations start with good intentions and more and more you see people trying to uh present themselves as the opposite or the counterveiling force against big tech empires and they start with hey we're not not going to have ads we're not going to censor you we're not going to kick you out from our platform but fundamentally those are just promises their air, fugazi, whatever you want to call it. And if there are no guarantees that they'll stick with what they told you, well, you can be pretty sure that at some point eventually money will uh will take the forefront of their thinking. So yeah,

don't become like Anakin. Uh yeah, the references are deep in this uh in this uh talk. So one last very important point before I go to the second part of the the talk which is much more on the subversive technologies and how they enable those safe communities I talk about is this people have their routines and they just don't want to change them. Uh this is a quote by some uh Harvard Business Review uh dude that said that attitude also persists as long as the organization permits it. So, it's super important, especially in crypto, not to be naive about those reasons. Understand that

people have routines and they're not going to just switch because they're putting their values first and then that those organizations need to be convinced in a significantly uh in a significant manner. Maybe the best example for that is uh a 40-year-old technology software that's proprietary, closed source that every business in the world depends on. uh and uh nobody has really made a move away from uh which is kind of catastrophic. I don't need to go into the details of it, but there is a nice Forbes article uh that's titled something like Excel is the most dangerous technology in the world. If you want to look a bit deeper into it, uh go ahead. So again to repeat my main

points, uh give organizations a perfect reasons to change their routine. give developers a perfect stack to develop healthier alternatives and give people a simple the simplest path possible to digital freedom. What does it look like? Sorry for the memes. It's going to accelerate now. Uh it means reimagining how we do a lot of things. Uh I use that example that refers back to what we're building at Fiverse at Fiverr. But the core point here is that for any uh social scientists in the room, it's uh the ability to imagine new affordances with our technologies because right now we're very constrained by a very

centralized model of computing and software and the moment we open that up you'll see that there are a lot of new design spaces that can be exploited. So why now? Why am I uh talking about a big switch and that this is a perfect moment to actually not continue building cloud apps and be competitive with those uh those uh trillion dollar companies. Well, one thing to point out is that we have hardness. This is a quote by Josh Stark uh which goes, we can give Ethereum code and receive a very strong guarantee that it will run anytime we call on it. We can give Ethereum data and know with certainty it will still be

available in the future unaltered and uncorrupted. And finally, hardness lets us control the future in small ways necessary to enable large scale coordination. This is what I mean by new affordances is new ways of even thinking about what we can do online. And this is what we're obsessed with at least uh in our team. I believe uh quite generally in the industry. So what does that enable us to do? It enable us to actually smash those centralized cloud empires.

Uh specifically what do I mean? First thing is technical counterpositioning. Um it's I forgot to put that slide in but uh until 2010 it's something I really like reminding people. Most of the internet traffic was peer-to-peer. A third of it actually was peer-to-peer enabled by Bit Torrent. And 10 years later, it's mostly uh uh cloud computing. Why? Not because those peer-to-peer networks were broken. They were just missing some very important primitives which we believe have been complemented finally by blockchains or how we like to call them community computers. It's a much easier way to uh

to communicate it to a bunch of academics that you're not talking about blockchain. You're just talking about community group computing blah blah blah. And people are get a little bit more uh enthusiastic about the idea and the possibilities. So I'm not going to go over the whole list, but peer-to-peer networks were scalable, hard to attack, easy to recover, easy to join, easy to exit, self-configurable AF. And on the other side, we have community computers that now give us strict global consensus. Opens the door for a bunch of decentralized identity. They have credible neutrality, computational and data integrity, super important. And my favorite, interup composability. And maybe the best place to look at that in

a practical way today is on Farcaster where you have a bunch of smart contracts interacting with each other under the same uh user interface but people are not even aware that actually it's different projects that have never talked to each other that are sharing the same daily active users because they they're enabling a certain mesh or stack of smart contract. Now, this is all good and well. We have a new technological stack. Is that enough? Uh, no. It's something that people are starting to pick up on much more or be more vocal on in in crypto, which is we really need to focus on UX. Uh, and I think that we're most in that left

panel. We think that super cool. What do you mean you have a privacy preserving ZK sauce with a little bit of privacy and all that? It it needs to be understandable to the end user. And currently we're doing a really terrible job at that. When I talk to people about peer-to-peer, this is a bit what they're imagining. It's okay. You need to uh sit next to someone, touch finger, connect one cable to their brain, etc. So, it's really not something that is intuitive. So it's really on us to make that UX understandable. Now for practical examples, those are two. Um uh one is Radical. Uh they just released a new desktop app

which is quite beautiful. I highly recommend you try it out if you want a peer-to-peer alternative to GitHub. This is really the team and application that has gone above and beyond to enable good UX that uh everyday developers will not feel really overwhelmed by in trying uh uh technologies that protect their data uh by design. The other is what we're building. Uh this is just a a quick screenshot of ducks. And my main point, the reason why I'm putting it here is uh not really to shill, but just to tell you that there is competitive advantages to having a decentralized stack. Google

is a trillion dollar company and Google Docs doesn't work when your Wi-Fi is not good. When your Wi-Fi goes off, you cannot write anymore on your Google Docs. It's a trillion dollar company and they force you to install an extension in in order to let you uh do some of the offline work you might need. And this is fundamentally connected to the way they've architecture their products which always require communication with some cloud in California which is not necessary anymore. Our devices are so powerful a lot can be right there on your local device. This is also what Radical does. Another UX advantage just

to finish on that is you can bring your own account. You don't need to be logged in by some provider even if they speak nicely and tell you that oh we're going to protect your data. No, bring your own account. It's not necessary. Um okay, five minutes. Uh just a quick one for those that are 3 minutes. Okay. Uh uh just to make it very practical. Uh this is if we ever turn evil or we're bought by uh whatever uh whatever entity, you can still use the app uh get all your data, decrypt it, get access to your account, etc. And this is fundamentally connected to the way that decentralized applications are built. final uh final

parts are the uh social counterpositioning uh um complementaryary to the technical counterpositioning and okay I don't need to explain the meme but let let's explode the train um just quick bullet points on what social counterpositioning is to take UX very seriously in order to rival the dominant apps out there is to support privacy local first and decentralized movement ments there are more and more and they are walking hand inhand with third point data protection regulations there is this uh weird myth that regulations is really the big enemy of

crypto it's not especially when we talk about data protection the GDPR is really the best gift that we've gotten as developers that value data protection and privacy and we should capitalize on on that. Uh just a quick anecdote uh uh 5 years ago Google was paying around 50k to the French data protection authority for uh their data protection violations. Uh that was an operation operational cost that they were so happy to pay and now it's $5 billion. So it's an existential threat and they cannot move as easily away from their architecture. Uh last few points uh offer by design data protection,

accessibility reliability interoperability, uh build in the open and grow through people's contributions. Do not lock in people or advance predatory business practices and offer radically new collaborative capabilities and experiences. Uh this two last slides, perfect 30 seconds. Uh be unapologetic. Uh I'm uh I'm being that a little bit at the moment, but you see somebody depressed in the bus, go with your phone, show them your app. Uh make uh make it a little bit awkward, but get the word out there about what you're building and new kinds of applications that are fighting

the good fight, let's say. And just to leave you on a comprehensive note, those are some of my favorite uh subversive apps in our industry. Uh lots others uh um outside, but really those are projects that I consider doing um uh the the good work, the good fight where promises are not the everything that you get. They actually show you guarantees of what they promise and it's quite important. Uh voila. If anybody wants to talk more, I have some floppy disc with me that give you access to some of those subversive technologies. Come chat.

Thank you so much. Maybe one quick question from the audience. One quick uh hi. Um what are you hearing from your users or from the community in general about performance of uh sorry about performance performance? Yeah. So file is built on IPFS, right? Uh yeah. Yes. Yeah. Yeah. Uh the trick is to also be local as much as possible. Like I said, our devices are a million times more powerful than they were 10 years ago where everybody was forced if they wanted to do collaboration to go somewhere else. And uh I can speak about

our users but more generally I think that there are clear UX advantages including in performance. If you look at uh your average app they'll have to load a page be redirected a bunch of times. Whatever you click on Twitter goes first through their t domain.x domain now and it slows down a lot of your interactions online. And maybe one good example is the work that Brave does uh where they show you in the homepage that hey you've saved 17 hours of your life by not constantly loading all the ad apparatus or uh communicating with a bunch of servers constantly without you even

wanting to. So performance- wise you see quite a bit of speed acceleration in software that is not built in such a centralized and cloud dependent way. Um We do have a team. We do monitor the performance of IPFS and I think it's it's pretty good for what it is. I just wanted to see if your users have more insights or complaints or improve this or that. Uh fundamentally we had to build a middleware we call it uh on top of IPFS in order to make IPFS quite usable. uh so yes IPFS is quite performant but I don't think the network was built with end users and applications in mind and you can see at the moment the discussions happening

that are trying to uh steer it in that direction but it came down to app developers to actually make it useful and performant and you know not expose users to uh so much complexity