On The Brink with Castle Island - Andreas Tsamados (Fileverse) on Decentralized File Sharing (EP.488)

Episode Date: December 20, 2023

Andreas Tsamados, the founder of Fileverse joins the show. In this episode we discuss: Andreas' academic background and path that led him to starting Fileverse An overview of Fileverse and the ambitions of the project. Why a decentralized file sharing platform is necessary and better than a centralized option for certain customer types. SAFE multi-signs The state of decentralized file storage. Token models and how these could evolve over time. The Fileverse roadmap in the coming year. To learn more about Fileverse visit fileverse.io  

Today on the podcast, I sat down with Andreas Somatos, the founder of FileVers, a decentralized file sharing platform for Web3 communities. Andres has a fascinating journey and he has fresh perspectives on the crypto space, so I was excited to hear more about his plans with this non-financial use case for blockchain technology. I think you'll enjoy this one. So without further ado, here's my conversation with Andreas from Fileverse. Matt Walsh and Nick Carter are partners at Castle Island Ventures. All of these expressed by them or the guests on this podcast are solely their opinions and do not reflect the opinions of Castle Island Ventures. Guest and host may maintain positions in the assets discussed in this podcast.

You should not treat any opinion expressed by anyone on this podcast as a specific inducement to make a particular investment or follow a particular strategy, but only as an expression of their personal opinion. This podcast is for informational purposes only. Brought down by bad mortgage investments, Lehman, which has 25,000 employees, will be liquidated. The federal government loans American International Group, AIG, $85 billion. This is a different kind of market, and the Fed is asleep. The federal government is stepping it to stabilize Fannie Mae and Freddie Mac, the two mortgage giants that have been threatened by the housing crisis. The Bank of England has pumped 75 billion pounds more into Britain's ailing economy with a new round of quantitative easing.

You print a couple trillion dollars and all of a sudden people start to worry. So out of this worry, we have something called the Bitcoin. Bitcoin. Well, Andreas, thanks so much for joining today on the podcast. Excited to dive into Fileverse. Great to be here, ma'am. Thank you. Why don't we just start off with a little bit of your background in, cryptography and what got you into starting file verse in the first place? So actually my background is more AI research.

I've done my master's and currently finishing my PhD on that. So it's on AI and cybersecurity, the University of Oxford. Specifically, the thesis is on human control of AI in computer network exploitations and attacks. Despite that, and you understand maybe the contrast, crypto has been a, long-running thread in my life. I've been in it since 2016, more focused on commons-based peer production, so from an academic perspective, and ironically enough, in AI already at that time, simply because smart contracts for me made immediate sense with

AI at the time, so I was wondering who was doing what in that. So I started working a little bit in that area. In 2019, I founded the crypto AI company that quite well and raise a lot of money and then I went to do my PhD. Now, the contrast between the two for me is clear, but to others, it might sound a bit provocative. I really see AI standing on one extreme, which is mainly today is used as a tool of surveillance, mass data gushing, analysis, processing, capturing your attention, manipulating it, etc. And it's probably the best friend of authoritarian regimes. A bit of a dark painting of AI,

but for some good reasons, or we don't hear it enough, I would say. And on the other side, you have crypto. Crypto that is all about self-sovereignty and letting people impose the terms of their interactions with the online world through cryptographic means. So this is what brought me to Favours. And specifically, the most tangible thing is

I got together with two guys, one called Vijay, based in India, who's been in crypto for a very long time. He's a tech genius of our team, really someone who thinks 20 years ahead. And Constantine, who's also in crypto for a very long time, probably had some of the most successful crypto campaigns and traditional media and exposure to people.

So I think some of his most successful crypto campaigns were in India, I reached hundreds of millions of people. So we got together, and it was the right amount of domain expertise, and excitement that brought us to actually wanting to solve what we consider one of the remaining big issues with blockchains and crypto, which is centralization in collaboration and in communication. That's fascinating. That intersection of artificial intelligence and blockchains seems like it's a very hot area right now. Obviously, you've been in it for probably longer than anyone, but what do you

see as the most promising intersection points there between the two technologies? I see, first of all, the angle of one acting as a good mitigation mechanism to the power of the other, like I described, for crypto allowing people to actually impose certain conditions to the inferences, for example, the model will be doing to the preservation of privacy in any type of interactions that requires a chatbot. And these things, at least a lot of those primitives for real ownership over data, real privacy guarantees through formal proofs are coming from research that at least I recognize more into crypto. So I think this is one important overlap. A second one that is, let's say, more consumer facing is that one of the biggest issues of crypto is the U.S. is the understandability. a high barrier to M3 from every single angle. And I think that's a lot of initial applications of LLMs in crypto will be to help transcend that terrible technical barrier. Yeah, that makes a ton of sense. So maybe transitioning a little bit to Fileverse, how would you describe the project?

And what was the ambition of the initial launch here in terms of what you're building? Great question. I'll start with the pretty title. The pretty title is that we're building a a trustless collaboration platform and workspace. So under the hood, what this means is that we're trying to bring back data control, privacy to the hands of users. And it's usually understood better by people when you contrast it with existing solutions. So there are two audiences that are best used to describe what Favis is addressing. The first is crypto people, people that are in Dow's, projects in crypto, etc.

Every single project that I know of and person I know of, despite the values that they have, crypto values, let's call them, are using Google workspace, Notion, et cetera, for the everyday interactions, but more importantly, for content that is meant to be public and meant to be here in 20 years. So this is technical documentation. It's maybe the bylaws of your Dow that you would like to have there forever, even if your group dispense or anything like that. And it's a point of centralization. It's something where one bad actor could decimate decades of crypto research and discussion. And we wanted to address that.

So our first audience, crypto natives that know that this is a problem, but haven't seen the solutions yet proposed. And then there is a bigger audience, the world internet users, that can be segmented as well further. Privacy first, people, etc. That experienced problems with the solutions that I talked to about. Only four days ago, I think, there's been a big scandal with Google Drive. A bunch of users have lost all their data.

Months of data erased, no response, no visibility over what happened, no easy recovery mechanism, no recovery mechanism at all. And those examples accumulate. They're still not enough to make a big wave of transition to privacy-first apps, but we think that it's the right moment to provide that to consumers. And we first start with crypto, but we're growing also to a larger audience. It's been fascinating to just see how crypto projects have evolved over time. Even something as decentralized as Bitcoin does have these centralization vectors.

I remember when I think it was Gavin and Driesen's ability. to commit to the repo was revoked, for instance. So it is fascinating to just see all these projects and then some of the centralized backends that they're actually using to work on the project. Absolutely. When you think about the core benefits of having a decentralized platform, this is more important to certain actors than others. But for those who might not be in the crypto every day in a Dow, how would you describe the real benefits of having something that is decentralized versus something like a Notion or a Google workspace? Most important point I would like to communicate is that nobody should even look at decentralized

solutions or will not look if they have bad U.S. This is with the premise of the decentralized solutions have a U.S. just as smooth and enjoyable as what Notion gives us. This is an acceptable premise and this is what we're betting a lot of our project on. We think that we have that U.S. In the background, what you benefit from, from this architecture, this security by design that we like to call it, is a few things. First of all, there are endemic issues in the existing centralized products like privacy invasion. The goal of a product that is free in the architecture of the internet today is you are the product, your data is a product, the processing over your data is a product, and this is why it's free.

So privacy invasions are given. Server outages happen regularly nowadays. And they come also with data laws, like the Google example of a few days ago. There is censorship, and this is a larger category of arbitrary decisions. A company goes busts. Rogue employee decides to take some of the data of 200 million users. All that is possible with a centralized architecture. And in fact, it's almost a notch towards.

coming out at some point. So the privacy invasions, there are always going to be scandals as long as we stay with this. I'll describe the architecture as server client model, which most of the internet has become. And the contrast is file verse

and is the client-to-client or peer-to-peer world. And just a bit of history, just because I know that people consider peer-to-peer a little bit idealistic, until 2009, maybe 2010, a third of all internet traffic was just peer-to-peer file sharing through BitTorek. That's not that long ago.

That's 13 years ago. And it is a very predatory model that brought us to what we have today, a predatory model with very pretty face, meaning great UX, amazing customer service if you want. But it has those endemic problems, including lock-in mechanism. That makes sense. I'd love to dig in a little bit on how it actually works. I saw a great talk that you did, I believe, at the SAFE conference talking about the role of multi-sigs. So maybe for the uninitiated, could you talk a little bit about safe multi-sigs, what they are? And then we'd love to dig in in terms of how you're actually using some of these underlying technologies. Absolutely. So let me first describe multisigs. Multi-sigs, maybe the best angle to describe it is that it's all of a particular. problem. The problem is that around $140 billion today is estimated to have been locked forever away in those blockchains because of bad U.X, bad key management, loss of keys, loss of

basically what allows you to get your money from your account. And multi-sticks are a solution to that, but also a solution to all the similar types of issues. So multi-sakes take us from a world where you have one address, one seat phrase that secures your unique access to that address, and it takes us to a world where you can have multiple addresses, all contributing or all controlling to some extent, to some threshold extent, the control of a smart contract that holds assets. So now, instead of having all your assets in one address, you have assets in one address that is a smart contract, and that smart contract is shared between different people that might not even know each other so that they can have material impact on the world by having cryptographic guarantees

saying, okay, since five of us out of 10 got together, the common money can be used in this situation. In that context, I'm curious how FileVERS is architected. Is it that there is a global multi-sig network for the entire platform, or would individual companies, per se? say, be able to set up their own configurations. Maybe you could talk to that. It's a great question. So the way we envision multisigs, and it's the case actually already now with our beta,

is that we'll take you from the trustless management of assets. So there is, I think, $60 billion that is secured today through multisigs. It's people or groups organizing and sharing control over common assets. And we're saying that instead of just assets, you can use the same logic to secure data. So every single user that we have can deploy their, if you won't call it, decentralized notion or a decentralized Dropbox or a decentralized Google drive,

whatever you need to make the easy analogy. And that control is extremely secure. It can be shared between people. It means that a group of strangers can set up a website together. they can manage contracts, contract signing, they can have a common branding repository that they contribute to together, etc. So this is the direction that we've taken with fibers and it allows us to bring some features that you don't have today with even Google or Notion, which are things like social recovery. You lose access to your account. You can rely on three people, two people that you

trust to recover your account for you. That makes sense. And then how does the initial setup work for that? You talk about user experience. Obviously, that's been a barrier historically to broader adoption for this technology, but how does it actually work to just get onboarded with a project? Absolutely. So behind maybe the more appealing side of followers to the average investor from the interactions

that I've had is that they realize that. that the barrier to entry into crypto through fibers is almost nonexistent, meaning today the most popular tools, including Uniswap, maybe AVE, maybe OpenC, require you to have some money into your crypto wallet. So you need to go through a centralized exchange first. We don't think that it is a serious way of getting crypto in the hands of billions. And we think that getting people to just interact with the whole crypto world without even either needing a wallet or needing money on the wallet is one way of helping people understand the value of crypto outside of gambling aspect.

Today you can arrive on our website. You can either share files directly without ever requiring a wallet or you can do something more complex like set up a multi-sig, have secure access over a storage, place and create documents, create whiteboards, create notion-like knowledge bases, all that without ever requiring either more than one signature, ever requiring to pay gas costs, etc. That makes a ton of sense. Now, obviously, storage is a piece of the puzzle here in terms of having large amounts of data on a platform like this.

So we'd be curious how you think about what that landscape looks like and what the interactions are between Fileverse and Decentralized File Storage Networks. Another great question is essential to what we're building. So I want to start by saying that we're building on the shoulders of giants. In the past decade, you've seen a lot of projects really try to bootstrap peer-to-peer storage networks and content addressing networks. And you have RWeave and IPFS that have come out of that. I think IPF has 200,000 nodes. Correct me on that, if I'm wrong, but they are successful networks.

but they're really not designed to be used by the average person. So we come in and we say, we don't want a centralized server to be a point of dependence for any user that wants to do daily activity we all do, create a document, share a file, upload a file, make a sheet, create slides, share them to only some person and not other people, and then revoke that access. All that, we can leverage,

those peer-to-peer networks to make it happen. So we use indiscriminately, actually, both networks, and we can add even more to them. Why? Because we're trying to push costs as low as possible for our users. And when you have open networks that compete with each other and where participation is permissionless, what do you have, which is an interesting economic phenomenon

that monopolies like Google Workspace or Microsoft Office don't enjoy, competition in price, if offering of UX, et cetera. So we have costs that are lower than storing your files on Google Drive, and we're trying to make the perfect UX around it, including creation of files, management, access permissions, etc. It's been really interesting for me to look at that decentralized file storage category, and particularly in light of some of the new technologies that have come out around things like Ordnals on Bitcoin and some of the zero-knowledge,

systems that are going to launch in the next couple of years. Do you believe that there will be more of these decentralized storage networks that you would plug into over time? I'm just curious what your views are on other projects that might emerge other than IPFS and RWeave as reliable alternatives. I think that already you can see with IPFS and RWeave, they have a different type of offering that they're trying to nail their narrative around. RWeaf is much more permanent storage. The phenomenon around Ordinoles tells me that you have a premium on a peer-to-peer storage place with very little space for you to store anything and that is super expensive. So I don't see it as being helpful necessarily for our average user, but for a news agency in Hong Kong that sees the government changing pace on the freedom of expression. or liberty of press might want to make sure that nothing will happen to that important

piece of data and pay the premium to have it on a network where they know even state actor will have a lot of trouble taking them. So there are different spaces emerging and I think that the moment you see them emerge with a little bit of traction, it's good indication that premium on Lindy, let's say, is an important factor. That makes sense. When you think about just the incentives of participants in a decentralized platform like this, obviously the threat vector would be if some of the actors start to turn malicious or start to be hostile towards the network. How have you thought about just building the platform with that adversarial mindset?

So, thankfully, we don't have to address that question from the ground up without any help. IPFS and R, we have resolved a lot of those questions. We think that some things can be improved. So incentive models have been defined quite well and they seem to be working. They can be improved or fine tuned further in order to maybe make it easy for any average user of workspace like ours to become also a participant in the network for something that an average computer doesn't even consider a light load on the browser and participate in a light fashion and the availability of data on the network and potentially be rewarded

through that ephemorous availability. This is something that Brave browser is experimenting with. They've had quite a bit of success. It's something that is on top of our mind, but it is also something that I'll be completely honest with you. I do see some non-incentivized dynamic emerging that may sound naive to the person outside looking in, but that the history of the Internet tells us to some extent sustainable. So Linux, Apache, Tor, a lot of those public goods have sustained their activity and even grown through something that cannot be said to be well defined in terms of incentives. So there is an aspect of that that we don't want to immediately kill by designing a,

good incentive model. We want to see also what organic dynamics emerge in the trustless platform that we create. That's a good dovetail into tokens. And you brought up Linux. Obviously, that's a very valuable product that did not drive value to either a token or equity, really, for that matter. It was really the company's building on top of it that captured the lion's share of value. How do you think about the role of a token in a decentralized network, whether it's necessary or not? Okay, so whether it's necessary or not, I would say it's not necessary. I think there are good examples out there that confirm that sense of mind. You have a lot of peer-to-peer networks, including BitTorrent before it was acquired,

that were great without a token. But that is not to say that tokens do not solve some issues that exist in those open networks and communities. Specifically, there's been a tension, at least in the, literature on open source software and commons-based peer production. So the usual academics from Harvard, including Benckler, the guy who wrote the Bazaar model, Eric Raymond, and others like that, kept pointing to the attention of it seems to be working well. People seem to be coordinating in a manner that is productive, but there are tensions that emerge. And there are people that leave. And there is a fight and there is a force.

and there are two communities that are built. They're weaker that way and they fizzle out within 10 years. Or the core community has been served by the SEC and everybody ran away. In those situations, you see that the token model does make sense. Because if you issue a token as a free open source software, for example, and you have already over 50,000 people using your software, companies participating, etc. You know that people might be actually willing to either participate more if they see

that there is a speculative token that goes up and down depending on the adoption of the software package. So it depends, but tokens have created a viable solution to the sustainability of open-truth software, just not for every time. And my suspicion is that if there were more regulatory, clarity on this issue, you'd actually see more tokens, where you'd see more equity-like value capture mechanisms built into products if there was just a cohesive regulatory framework where the developers themselves would not be at risk of an SEC action, for instance.

I completely agree. Actually, maybe this is controversial, but I think that ICOs were one of the best inventions of the crypto space. It was abandoned way too quickly. But I could definitely see that resurging if, for example, the European Commission hosts their current line and actually define it a little bit better and allow for these types of events to happen, especially in a way that ends up being better for both the buyer and the issuer of the ICO. For example, Daico model, which is the further milestones you complete, the further shares can be issued and bought by a community of whoever is willing. And this could be incredibly interesting.

It's much better than an IPO model. I agree with that. What Ethereum did is not replicable right now, at least in the United States. But if you got something like a safe harbor proposal, which Hester Purse at the SEC has advocated for, you could imagine a world where that funding mechanism comes back and actually fits within the regulatory regime and allows for this progressive decentralization of a particular network, such that the asset would not be a security when it's used in the product. Exactly. If a lot of the regulatory tension, at least in the U.S., is around the question of what is a security and what is not, maybe putting down a proper plan on how to avoid becoming a security through technical and social distribution means,

then you have something that is quite productive for everyone in both. Let's talk about what's next for Fileverse from a product perspective. What's ahead of you guys for the next year? I love the question. I just came out of a long protocol, actually, with my team. So a few things. Right now we're in AllowLis mode. So we're trying to test a few things, improve retention with a big group of core users that we currently have.

We're opening the doors from time to time in order to test how new users appreciate the UX that we've been developing. And right now our big focus is on making the best UX possible so that during the next hype cycle, for example, we can be that project that everybody is proud to show to their friends or their family during Christmas dinners. And no, you don't need to buy anything. No, you don't need to give your bank account details passport to Coinbase to even participate in my world. you can just click there and here you go. You have some better guarantees of control and privacy. So that UX, we want to nail it. This is why we're keeping also a bit closed.

We're working on a lot of fundamental improvements to the backend. So we have something that we're quite proud of, which is a way of authenticating and authorizing people, authorizing access to some data, in a way that is completely decentralized as well. So we're improving that, making it ready for the big release. The big focus is retention for the next few months until we do the big public launch. Right now we're focused on the crypto community, then the privacy first community,

people who use signal, et cetera. And then we really want to be able to show anyone outside that. That's it. Crypto is now. Onshane is the new online. you have the same you wax, you just enjoy better guarantees for your liberties. Well, I'm sure a lot of people in the audience will love to show some family members something that is non-speculative and not scam related in the crypto industry,

given what's happened in the past year? Before the public launch, what is the best way for people to learn more about Fileverse and to be in touch? So I'm going to give you a proper crypto Twitter answer, Twitter. So Twitter is really where we try to document on a regular basis, every change, every discussion, every interaction with the community, etc. So go to our Twitter. It's at FileVos, very easy.

And I'm also a fan of making people test products themselves. So anyone from your listeners that would like to try out FileVos can drop either DM and said that they're listeners of your podcast or directly to our FileVos account or comment on their video and we'll make sure that we'll make sure that we'll. whitelize their ENS or their zero-x address. That's awesome. Well, Andreas, this is fascinating. I'm really excited about what you're building and excited to track it through the next year. So thanks for joining us today on the podcast. Thanks so much, Matt. Thanks for listening. Thanks for the amazing questions and taking time with a non-financial application in crypto.

My pleasure. Thanks for listening to another episode of On the Brink with Castle Island. To find out more about Castle Island, visit castle island.V.C. To listen to all of our podcast episodes, please go to On the Brink-Podcast.com or just click on the tab in our website. Thanks for listening.