“The Next Privacy Revolution Starts with Good UX" Andreas Tsamados x Fileverse

Hello everyone. Uh my name is Andreas Muryato online. I'm one of the co-founders of Fiverr. We're the ones building ducks. D sheets.

Super happy to be here.

So Andreas, hi.

Thanks for coming. Pleasure.

So for the context, we are Devcon in Berlin. It's kind of already finished 6 something

and we have this like late late latest latest talk for today. We had a lot. Um so thanks for joining. Can you please uh tell for the people who don't know you a a bit of your like backstory? How?

So first like what do you do now and like how do you end up there like just few minutes uh your like professional maybe personal like journey towards

okay yeah uh I'll try to keep it a bit short uh let me get it out of the way already that uh I'm currently finishing my PhD uh so it's something that uh I'm very keen on uh finishing soon because I want to focus completely on uh fverse But my PhD at the moment is uh um on AI and cyber security generally it's um specifically on the use and control of AI systems in uh hacking operations between states uh specifically so how APS are advanced persistence threads are actually using uh those new uh foundation models in order to uh hack uh uh critical systems. So that's what I do on the academic side and uh most of my time goes into FER which is a company I'm a co-founder of and uh we've built tools like ddxnew or dsh sheets new uh which are privacy first and decentralized alternative to the Google suite. So Google Docs uh Google Sheets, Excel, etc. um how did I get here or how did we start that it's maybe not the best thing to uh to say on the video but uh it's mainly anger uh anger spite that drives us at least me personally it's um uh you know everybody has their own thing that are powering uh their work and why they're taking directions certain directions in their lives and there's uh kinds of injustices that I uh absolutely do not have patience for or do not want to accept and I don't want to stay idle. So um this is what uh led me and my co-founders to uh creating fathers

which is uh we fundamentally believe that people care about privacy. Don't listen to all those naysayers outside that say, "Oh, people don't care about privacy. You need to force it into them." Blah, blah, blah. No, everybody cares about privacy.

It's just in a world where all your interactions are digital. You cannot expect people, especially, it's a very privileged point of view, to think that, yeah, they're going to take the extra steps in order to secure their own privacy. Even if they know how bad it is for them to be super transparent online, they need to have a smooth experience and a smooth transition from mass surveillance applications that are the main ones powering the world right now into good alternatives. So our big purpose is to create UX that is absolutely competitive to what you have out there and uh to drive people into uh uh privacy and data protection by design.

Mhm.

Well, that's awesome. Actually, my personal kudos to you for UX. I mean, I really use the your tools a lot and they're super cool and that's something which is for me like being in web three space since I don't know 8 years now. I it's really hard to find a good UX. Why why you managed to do that and why so many people struggle to do that?

Uh in web in crypto specifically or

Yeah. Yeah. In web three. I mean web two people can imagine the importance of

um in crypto it's uh there are many reasons I think that uh that explain it but um not to try to be too apologetic but it is true that until quite recently the the crypto stack was not the easiest to handle. Uh so a lot of people had the expectation that uh okay people want to make money. Most applications are about money making. They are financial. We're the only naive stupid people that 3 years ago came in and said hey non-financial applications in crypto is possible.

So their care for UX was maybe not as important because they knew that there was a much stronger incentive for the users which was I'll sign as many things as you want as long as I have 15% APY on my uh on my pool or on my staking.

So it wasn't that important for developers. uh then the stack was also very difficult uh to uh uh to build with and now I think that it's a mix of complacency and maybe um it's quite sad to say that but maybe not talking to your customers enough. Uh if you follow us online you probably know how we are. Uh we're very open builders. We we take in every feedback request etc.

we respond to everyone. So this creates an environment where people users know that they can speak openly to us and they know that within a few weeks they'll see those changes happen on the app side and this is why also we have such a strong community that okay has understood that those guys are really trying to have good UX. One last point, uh maybe something more general for later, but I fundamentally believe that the UX of decentralized apps can be much more competitive uh to cloud-based apps. I like to give two examples which I gave at my keynote which was um Google Docs powered by a trillion dollar company uh doesn't allow you to continue writing on your docs if Wi-Fi is not good or Wi-Fi went off. It's a trillion dollar company that's been around for 20 years.

Catastrophic. It doesn't make sense. And they force you to install an extension, etc. This comes from, you know, a very, it's a business model decision. It's not a UX necessarily decision.

And you see where the constraints are for those companies and how we can be competitive because our app doesn't need to constantly talk to a server for anything and we actually leverage the very strong computational power and storage of your device.

We're like, okay, we don't need a server to be doing everything. most of it can happen locally on your computer. So this is why offline works without you downloading anything on flowers uh on ddogs new sheets. And the other is um if you do split screen

uh on Google Docs, you need to scroll horizontally in order to reframe the editor.

Oh,

which is catastrophic. It could happen automatically, which is what we do. or another I'll finish with that is uh you don't need to force everyone to create an account.

I want to collaborate with someone. I don't want them to be forced to create an account or log in with their stuff. No, it can be less predatory and you can let people use your stuff without forcing them into a locked in account.

Okay. Um maybe um you could share a bit of your challenges that you had like making up uh the docs and files because there are so many builders in the space and then

um they have so many problems like starting like what to build, how to build, where to get money. So um maybe you can share some of the challenges that you had on your builder's path and how did you overcome that? So there would be more people in space like building something cool and it's called URX. [laughter] I hope so.

Yeah. Yeah.

Why it's just you? Uh well there are many aspects to my answer. Maybe I'll focus on uh the first which is that we need to be realistic with even young builders. um they cannot subsist or sustain themselves out of purely being driven by their values. They need to see money coming in.

So this is probably one of the most tangible uh limitations or constraints on builders and which forces them into making compromises on a bunch of things. So I'm going to compromise on um the smoothness of my signup process. I'm going to compromise on um the end to end encryption because it's difficult to implement and it's not as easy as just server side encryption. So that difficulty forces a builder to make compromises and if you're on top of that constrained by the money that comes in into your organization or lets you actually build then you're more likely to go into those compromises. you might have the best values ever.

Uh you need to be realistic. You need to feed your family, etc. So, it's understandable. And maybe something that I'm on a positive note, something I'm quite hopeful about with crypto uh with the web 3 space more generally is that there are a lot of opportunities for getting donations from people. This is how we started.

We had over 10,000 donors. it gave us so much life encouragement and also a bit of financial backing.

Um then you have millions of grants that are available and up for grabs in crypto. So all those things can contribute to making the life of a builder um better and ensure that they won't compromise at the first big obstacle uh in their in their building processes. Uh and you see that more and more in hackathons, people are starting to uh um because the crypto market is going up, people are becoming a bit more ambitious about their ideas and um what they can actually build because if uh

this big uh entity that raised 150 million, made a token that went to 1 billion says, "Hey, we have $50 million up for graphs for builders." You can think a bit bigger. So this is quite practical and uh it's probably one of the first requirements for a builder which is okay secure yourself a bit otherwise you're going to compromise on everything that you wanted um to to be different at

yeah jump into your PhD is like a super super relevant topic for nowadays so it's like we are in June 2025

uh so your hot take on a recent blackout in Portugal and Spain a few minutes ago and in terms of France as well.

Yeah. Uh if I have a hot take on it.

Yeah. [gasps]

Look, I'm going to keep my conspiratorial opinion to myself because I don't want it to uh to be overindexed by people. But um to me it it reminds us all of something very important which is our most core infrastructure for communication, collaboration, human coordination in general um is quite centralized.

Yeah. Um I was showing during my talk a picture of all the submarine fiber optic cables uh that uh power the internet and how in the past five years uh it went from consortia from non-governmental organizations with some governments getting together blah blah blah uh to private companies Google Microsoft laying their own infrastructure for the internet and creating those super significant choke points. So finally the Europeans experienced it a little bit but you talk to someone that is in Madagascar or somebody that's in Bangladesh, they've been experiencing it much more because the the um the concentration of internet submarine cables in those regions is uh much smaller. So it is uh much more of a regular thing to see internet being unaccessible to everyone. So first thing I'm happy it happened a little bit uh in Europe.

A good reminder. And second, I'm going to plug in a project that I loved. Um it's called Guifiet.

Um it is the biggest I think mesh network in the world. It's in Barcelona. Uh it's volunteers that installed on top of buildings, on top of towers, small hardware

that has uh enough of a reach for the town to be connected and have an alternative

system of communication on a local level.

So when internet goes down in that region, you know that Barcelona has a backup

Wow. uh uh service that is completely volunteer based we find same thing where I come from in Greece in Athens

uh it's not as big as Griffinet but people have been installing mesh networks uh hardware in order to have some communication within the city that is not dependent on some big providers

so that's my pick

Wow that's super cool I didn't know about that in Barcelona even though I live there that's nice to know okay Yeah, go visit them. It's really nice.

I will explore. But like when it happened, I was with Michael actually in Madrid and we actually got push notification and then a friend of mine told me that it was through the Bluetooth system of Ho that is like the air tech uses actually because I couldn't access anything but then I got some pushes and it was a bit weird.

Um yeah, so okay and being maybe less specific. So your PhD is like general it's not like specific case uh I mean it's it's not concentrated on some specific case it's just on like how how AI is used for cyber war let's say

uh yeah yeah so uh my very specific focus is uh on human machine teaming with foundation models for hacking operations and I'm looking a little bit more on the offensive side Mhm.

Um, so, um, are you familiar with the conference called Defcon, but with a F,

not with a V?

No.

Um, it's the biggest hacker conference in the world. It happens in Las Vegas once a year. Mhm.

It's where the NSA and uh and the other

uh the the rest of them recruit a lot as well. But um it's quite an amazing space with a lot of very knowledgeable technical people. And this is where for example I go in order to talk to uh to people and to hackers and get some operational information about their work which is lacking completely in academia. In academia, if you look at the field of cyber security, purely technical,

some policy stuff, some security studies, which is much more like war studies, international relations, and then when you try to look at, okay, what are people actually doing, what are their mitigation processes, what kind of framework are they following in order to have a good loop uh feedback loop between a human operator and uh agents, you see that there is very So this is what I focus on. I'm trying to uh uh illuminate a little bit that space and give people information about what's happening behind closed doors. And there is a very nice report uh by uh Chad GPT people. I don't even want to say their name because their name is just it forces us to characterize them in a way that doesn't characterize them. It's closed AI.

Anyways, um they have a report if you're interested on the latest APS that have been using foundation models for their hacking operations. It's quite quite interesting.

Okay. Okay. And um like speaking about current academia, um you said it's very technical, but is it contemporary? Is it like um progressive enough or it's a bit like behind what's going on in reality? uh always behind uh of course um but for good reasons.

Um there is a complaint that I hear a lot especially by American academics but also the new tech oligarchs around the world that keep uh telling people that oh academia is dead universities are not necessary anymore uh because they're too slow. It's so important to be slow in your research. I want to be able to have 5 years to explore the sex life of ants. It's important for somebody to spend that time in order to pick up on all the nuances and actually uh uh create good research. This is where you see also that the incentive

is intellectual. It's curiosity is scientific. It's not I got a grant, I need to ship a paper within 6 months. Whatever happens, my KPI is clear. I need to get it out.

So it is slow for good reasons and then it's a bit contradictory what I'm going to say but you also see some of the best bleeding edge innovation coming from academ because if you look at the AI field what's happened in the past 10 years you think it's companies that led that innovation no it's academics that for the past 50 years have been obsessed about deep learning uh models that have been obsessed about uh non-deterministic systems and how they can actually help us uh or uh yeah uh that so it's quite important to to always recognize that um those billion dollar newly minted companies are building on the back of academics that are poorly paid, poorly resourced, barely even acknowledged. Uh, and they're claiming all the fame and they're claiming that they understand what they're even building. I completely reject that premise. Those big new AI companies. Sorry, it's a moment where I get a bit passionate, but it's

big AI companies.

Go crush them. uh their great propagandist funion from the EU commission is reusing the same narrative and words that some Altman has been pushing onto the world for the past 3 years which talks about existential threat of AI and it's going to uh is going to kill humanity all that which is established in absolutely no sense of reality. Why? because they're selling a product and we're all following that narrative as if they knew what they were talking about rather than they know exactly what they want to sell to us. And it's problematic and academics are important to diffuse a bit that or rather demystify uh what those organizations are really selling to the world.

Oh, let's demystify AI killing humanity. Why is So uh are you a believer?

Um I would say I don't have a straight position around that. So I know this narrative.

I don't I mean I think yeah it sounds possible and we all have seen the I don't know when I was a kid I have seen Terminator 2. Okay. Yes goes into that narrative that is possible probably. Yes. Especially when AI takes a lot of control and then all the infrastructure is controlled by AI and then so many things are controlled.

Is it possible? Yes. Is it going to happen like super soon? I don't think so.

So for me, I leave a chance of a possibility for that. But I never dig dig like into that. That's why I'm curious uh to ask an expert in the field because I'm not.

Okay. So um I'm with you. I think there is maybe a possibility at some point but uh what I'm really rejecting hard is that we're any closer because of uh foundation models having been invented that uh new architectures being out is us getting so much closer to an existential Terminator like scenario. So they're playing on years of science fiction and what it has built in in our cultures and our perception of what AI might be

which by the way 2,500 years ago humanity was already freaking out about that. The stories of golems

uh which are uh entities that are independent and uh um can potentially take over humans etc. It's uh Old Testament uh type story. So it's quite insidious. That's my point. for those companies to play on this uh exist existential dread uh and um use the mysticism around very complex mathematics in order to tell people this is coming but also good luck determining if we're any closer to this this thing happening because

when it's super yeah when it's super complicated for anyone to look into it and be like, "Oh, yeah, no, I can determine that is completely bullshit." No, because they're hiding behind the wall of mathematics and statistics that um yeah, it's quite insidious. So, that's one thing. The other is that um those uh those systems do not reason like humans reason uh reason. Uh those systems are non-deterministic.

those systems uh cannot count how many RS are in strawberry.

And when you see very concrete examples of their limitations,

you start realizing that okay then how do we make the jump to those systems will develop a competitive sense with humans.

Through that competitive sense, they will want to eliminate some humans and eventually all of them. through their understanding of the world they'll be able to crash very important critical infrastructure etc. And it's a distraction from what? from the very real dangers of Google, Microsoft and all those uh big tech empires have forced everyone to suddenly have AI agents and LLM in my email, my docs, my drive, my all those things as if we had asked them as if we really wanted this So, first of all, I don't want it. Second and this is uh sorry I'm going to show a bit uh my research but I have two papers published on that which is how vulnerable uh AI systems are to really the the most basic type of hacking.

We've never had such easy attack vectors to people's personal data and identities uh than today because those big tech companies decided okay everybody needs an AI but those AIs are very easy to to to uh to hack into. I'll give you two examples. One is called indirect prompt injection.

Uh it's me sending you a document.

You open that document. Even if you don't open that document, Gemini pre-scans it. And I have some hidden uh ASKI um um

prompt

prompts uh yeah instructions for the LLM. It processes it and it sends me your whole history of email interactions.

Beautiful.

Okay.

Beautiful.

Amazing.

Yeah.

Another is uh through data set poisoning. It's called actually it's called model poisoning which is I know that uh most AI systems use the same data sets and they rank different types of um uh websites online for their importance in terms of data for the training of their models. So everybody goes on Reddit, everybody goes on Stack Overflow, GitHub etc. So I know that those things are the most important sources of data for the models. So what am I going to do?

I'm going to go onto those platforms. I'm going to start uh uh spreading some very specific information knowing that it will be captured by the data scraping or yeah I'm going to do that and uh I'm going to hide some hidden instructions or some really completely information and then I'll be able to do some really dangerous uh remote control or um uh um uh or have some exclusive understanding on how I can manipulate a model because I know that it was trained partly on poison data that I distributed online. And today we're probably seeing that. It's too early to tell, but I can bet my life on it that every big government is already starting to poison our most uh important websites in order to make sure that they have some hidden backdoor access to those models. and this man we we're completely rearchitecting information uh flows online through that and it's catastrophic and I want more people to to be aware.

Wow. Then there comes a question how do you use AI but do it consciously and securely and private as well because it's hard not to right I had this talk with Leeris from Rodki and he he said a wonderful thing that like AI it's not just I mean you compromise on your productivity because like he's making the same time coding and like making his presentation and if he doesn't use that and he's far far behind. So, it's really hard not to use. But then how you do that? Okay.

It's very hard not to use. We're 2 years into those products actually being usable and suddenly we're all pretending that we can't live our lives without them. Like, okay, first of all, I would like to push back a bit on that. Maybe I'm a little bit too much of a loite. Maybe I because I spend my time researching those systems.

pallet. I'm a little bit more reluctant to adopt them that easily.

Um, but how to use them privately? Um, sorry, I'm going to shield DOS for a second.

Yeah, go for it. We love you. We love you anyway. So, go go for it. Thank you.

Uh, so on DOS, we wanted to address that quite directly because a lot of people including Vitalik have been spamming us about okay, can we have AI because it helps us with our productivity. I understand. I I'm not going to tell people, hey, you know, you shouldn't use it. It's my personal take that you shouldn't overuse it. Um, but instead of forcing people to adopt like notion did, Google did, uh, everybody uh, literally did, which is we're going to push our AI with our server onto you and you cannot say no.

It's there now and it's going to process your files whether you like it or not. We tell people, hey, if you go in your settings, you can enable AI and that AI is not even run by us. It's only if you have a local model installed on your computer. You don't know how to do that very easily. Three clicks.

Now you have a local model. You pick whichever you want. All open source. And uh now we made the UI very adapted to you having connected your local LLM and you can have your model process your documents if you want if you explicitly ask uh it can have the context of those documents in order to give you a better output better responses better advice recommendations etc. It's native to the editor, so it can uh uh make its output in markdown, create tables, you know, it's uh directly connected to what we built and all that is private as Why?

Because there's no server. It's your device. And your device, after the past 15 years, our devices are so powerful. A lot of what we're doing online, we've moved to cloud. Not because um it was necessarily the right choice.

It's just that for collaboration and heavy tasks, our devices were not good enough at the time. Today, we've got rocket ships that we all have in our pockets and we should be using them so much more. Like imagine in the times of early pirate bay, early bit torrent, people having devices we have like like we have today. It would have thrived even more. So we want to remind people that hey you have everything you need locally.

Uh let us make it easy for you to stay private, stay uh protected uh stay secure.

Oh that's super cool that you have this option. Actually I didn't know that. So thanks for sharing and it's very useful. Yeah, because like yeah mostly people say yeah use for instance I don't know deepseek make it locally of course it's not that powerful it not that maybe accurate and up to date but it's yours then

yeah it's super powerful I really I consider deepseek so much more competitive than uh 03 or whatever uh that I need to give my data to. Um so yeah.

Okay.

And uh like in the very beginning you've said that you think that many people consider that privacy is a problem. Uh which mostly people that I talk to at least doesn't think so in the privacy bubble as well. So um why do you think so? So my point was specifically that I hear way too much nobody cares about privacy

and nobody cares about decentralization

every day on crypto Twitter and beyond you have the same being repeated to everyone. Sorry I'm using a lot of profanities. Um and I I really reject that premise completely. It's just you cannot expect people to you know spend that cognitive power in order to process all the hops they need to make in order to get a bit of privacy.

So uh yes of course people are going to abandon their privacy because everything structurally speaking is forcing them to abandon their privacy. So we need to be much more aware of that and we need to help people actually achieve good levels of data protection. Um and uh it's on us to take the burden of it not on the end user it's on the developers.

So yeah and for that you know one last note on this regulations are our friends.

H take wow that's a really hot take. Uh let's let's expand. Okay,

that sounds really really interesting. I mean um go go for it.

Okay, then um okay, two examples. Um the GDPR requires requires companies

to have data portability and if you don't have that uh you get fined. Unfortunately, enforcement is not really keeping up uh as it should. Uh instead, we got uh stupid cookie policies uh spamming uh everything we do online, which was a complete failure. But anyways, um and that data portability is something that should speak to every new app builder, especially privacy app builders or decentralized app builders because it gives us an opportunity to challenge the monopolies, challenge the the big tech empires that have either the power to destroy you or to buy you out in a way that you might be the you might be Uh, Mother Teresa, if somebody comes and says, "Hey, you want a billion dollars?"

Hey, your brain might start thinking, I can do much more with a billion dollars. So, uh, how do we fight back? How do we actually capture a significant segment of the market? It is by leveraging those laws that are speaking to us. I want to be able to say that uh I want to be able to sue Google and say hey there is no data portability here there is no interoperability with the other applications and you're violating the most fundamental digital rights of EU citizens so we do that we force Google maybe they start opening up their their their platform maybe they start adopting open standards and we can capture back some segment of uh that market.

That's one thing. Second is that uh beyond data portability you have it's a little anecdote but um Google used to pay around 80k a bit more to the kil the French data protection authority um for its violations of you know locking in users lack of privacy leaks online whatever Um, today it's become an existential threat those fines because there's not a 50k 80k uh fine which means nothing to Google. It's up to5 billion. And those companies that have built for the past 30 years their architecture around lockin mechanisms choking every possible data out of you selling it to thousands of data brokers. um they're starting to panic.

They're starting to see those laws as an existential threat.

And what generation of apps is actually built in many ways GDPR compliant by design is decentralized apps. It's an app like ours that says, "Hey, you don't want to create an account with us, don't create an account. Bring your own. Bring your email. Bring your wallet.

Bring your PGP keys. you we don't need to force you into uh predatory business practices and actually you can guarantee that you won't be forced into that. So really I I I think that it's finally an amazing opportunity for all of us builders with values uh to challenge those big empires. And just a final note on that actually I'll stop here.

Okay. [laughter]

Okay. Okay. Okay. Um then um I mean as we talked before it's really hard to compete with this big tech especially in terms of UX but uh maybe you can recommend some cool tech which is more private uh that people can use actually now.

Yeah.

Um like few examples that that are cool.

Yeah. Uh you want applications for end users or you want for developers?

It could be both. It could be both actually but your take on that. Yeah.

Uh okay. To replace Slack uh use or Discord use Cut. Amazing uh amazing team is uh written CWTCH I think. Um they used to be I think or very close to the tour team and they're doing amazing work. Uh you want to use an alternative to GitHub use radical.

xyz. You want to use an alternative to Google Docs, use GDOX new. An alternative to Sheets or Excel, use DS Sheets new. Um, you want to um you want to protect your internet browsing uh really without having to hop through million UX difficulties, use Brave. Brave is really, really amazing.

you want to use private AI while you browse the internet, use Brave Leo,

which is also really amazing tool. Um, what else? I really have a long list of those, so I can go all day. Uh, you want some help on the legal uh side and you're building good technologies, technologies that are interoperable by design. Uh, check out um check out datarits.

ng. uh they're in Berlin. I think they're Netherlands based, but they're really amazing. They litigate against governments, against big tech companies, and they're forcing them to uh apply data portability and can also help potentially companies get some money out of those lawsuits. Why?

Because uh they can say the small companies can claim that hey uh it was a anti-competitive practice. So, we lost a lot of money. So, Google owes us x amount of uh of cash.

Wow, that's really cool.

Yeah. So, uh well, a few practical things. Uh if

maybe a few more for builders.

For builders, uh for builders, I would recommend checking out UKANs. It's a user controlled authorization networks. It's really really cool. Um, I would recommend looking into VOP RF uh uh- ID. It's by the privacy and scaling team uh at the Ethereum Foundation.

They're doing really amazing work. Uh it's for uh privacy preserving authentication basically. Uh what else for builders? Uh try not to use cloud as much as possible. Peer-to-peer systems are out there.

I know it's difficult to use them. So some teams have been going above and beyond to make them useful. Um our team included we created a fus middleware for developers not to have to overthink okay all that. Uh what else for builders? Um just a general practice advice is try not to lock in your users really like I understand why you want to bump your retention, bump your growth, etc.

But it's going to bite you in the ass. Uh create that. Uh this is why we have a strong community at Fus where people know that hey we even f even if fus turns evil

they can continue using their account data app without uh without worry because they can download it out of IPFS and that's it. Everything works. So yeah those things

imagine we are 30 years from now

and we won. So we have the internet you dream of like

secure, private, decentralized whatever.

Hallelujah.

Uh can you describe it and how it's different and how is life different in this like techno paradise? O okay, I need to decide on a few dimensions otherwise I'm going to speak for hours and I have a bit of a I tend to blah blah a lot have you as you've noticed. Um so first of all um I can take I can be on my favorite social media uh there is a new competitor and I can immediately port to the new competitor but not just port and I know I have my PSP my name and that's it. No, I can take my whole social graph, who I'm friends with, uh who I connect the most with, what are the topics I'm the most interested in, etc. So, social graph,

last point was more about algorithmic independence, but I'll point that out as well. Um, I would be able to crossplatform communicate. It's not because I'm on uh like email, right? Uh, everybody uses SMTP protocol uh under the hood. I should be able with my uh uh with my Proton account uh to talk to somebody with a mailbox account, which is the case, but I want the same for social media platforms and our everyday apps.

I want my WhatsApp uh friends to be able to communicate with me on Signal. It's so important. It's so important. communication is really a base primitive for everything we're going to talk about for this new world. Another is algorithmic portability.

So the the especially in 30 years there's probably going to be much more AI and algorithmic moderation, content pushing, etc. curation. So I want to be able to define to my identity a few preferences of how I like to explore the web. I want whenever I do something on the web to have a way for any website, any algorithm to that interacts with me to not know who I am but know my preferences and a lot of noise in it so that they can never be anonymize who I am

and know that hey don't show me those kinds of ads. Hey, I don't want to hear about uh American politics. A I um I have a a tendency to believe humans rather than AI agent giving me a summary every time I I search engine some information. I don't want to see those things. So give me first Reddit results.

I trust a community of people talking with each other a million times more than any AI agent giving me a summary of whatever and they have a propensity for error that is that we have no idea how to get rid of. So you'll have this. you would have also uh you would have much less of what we see today which maybe I'll finish on that because it pisses me off a lot which is um there's a term academic term it's called ethics sharing s h i r k i n g which is a process of exporting practices that are illegal uh where you're uh you're based

to usually the global south, countries in the global south

in order to get data for example

and from that data you get massively cheaply and abusively from the global south. You enrich yourself incredibly in your own country because hey, you did the illegal stuff uh in wherever and you imported back the byproduct that made you a multi-billion dollar company. You know who does that

worldcoin?

Uhhuh.

And I I tell that story as much as I can to people. My co-founder who lives in India arrives one day in his uh city. I'm not going to say which city, but it's really not a big uh uh city nor a big tech hog, whatever. Arrives by train uh and sees a huge line, follows the line, doesn't know where it goes. At the end of the line, it was a little stand with a orb.

What? and paying people peanuts.

Wow.

In order to get their most their most personal biometric data and enrich themselves and then claim which everybody in crypto likes to repeat as well. Oh, we have 2 million daily active users. Who are those users? How much did you pay them? And how much did you abuse them for their conditions in order to enrich yourselves?

is complet completely unacceptable and every big AI company is doing the same. They have uh they have data labelers, curators, data cleaners in Bangladesh usually in Nigeria as well. They're paying them less than nothing. They're exposing them to really aggressive uh uh data and content. Same for moderators on Facebook or whatever.

they're exposed to stuff that nobody gives a what kind of trauma they come out with because there are no regulations there. So, I really hope we won't be seeing that anymore in that new world.

Uh yeah, let's add a quick notion for people that don't realize like the danger of like sharing your private data with World Coin like why it's so bad and for those people as well.

Yeah. The first point is when you see a company having such a predicatory practices where they abuse uh uh less fortunate people in order to enrich themselves and not give anything back to them. It should be the the first and maybe end point of my argument. That's it. You don't need much more.

Uh the rest is okay. It's there is nothing more personal to you than your biometric data. what you're composed of, what makes you unique at the biological level. So, you know, be careful with who you share it with, how you share it, what kind of guarantees you get around it. And don't forget that, okay, maybe this I'm 80% sure, so I'm still going to say it, but uh last time I checked, um, uh, that company, the Orb company, uh, considers biometric data that they capture their IP.

It's their IP.

Your most valuable personal information is their IP.

Wow.

Completely completely crazy. So, uh yeah. Uh be careful.

Well, well, yeah. And the final short one, uh since we are doing that for the academy, so it's like people who are kind of like to study and consume in a good way like um information. So can you recommend like a cultural piece, book, film, music, whatever that might enrich someone's life?

Okay. Wow. I have so much. [laughter] Okay. I'll pick the I'll pick different types of media formats.

So I'm sure somebody else did it said it in the interviews, but it's important you hear it from more people. Uh the internet's own boy. Amazing documentary. Free on YouTube. Go watch it.

It's about orange farts. Uh it's very important for those who care about privacy. There's a paper that came out maybe 20 years ago, but that I it came out in 2004 actually by a Yale uh legal scholar professor

um that is still very relevant. is called the two western conceptions of privacy and it contrast American versus European. Is it right? Is it a duty? Is it It's quite interesting.

Um uh this and maybe one more. I'm not going to show my papers. Um yeah, the the the NGO that I talked about, data.go, NGO they're really doing great work so I would recommend to go check them out and if you want community if you want people you can connect with there is a lot in Europe uh you have the chaos computer club you have weet you have uh so many more you have CBase here in Berlin find your people uh that's maybe the most important cultural insight I can share yes read those things yes watch those things but find your people find your community. Those are the people that will give you the energy, courage to continue what you're doing and not feel like you're the only crazy person in the world that cares about how we're treated online.

Wow, wonderful advice. Thank you so much.

My pleasure, man. And this floppy disc is for you. Write a little message on it.

Old tech.

Is it a real one?

It is a real one. So, on all of them, I've put NFC chip

and if you put your phone on it, it will redirect you to DO.

Wow. Oh, that's cool.

And uh in some of them, I compressed a version of Doom, the old video game. So, if you have a reader, you can check it out.

Wow, that's super cool. Thank you. That's I believe that not everyone watching this video would know what

Yeah. [laughter] Somebody made a good joke about it. Uh what was it? Ah, yeah. Uh on most editors, the save button is a floppy disc.

True. True.

And somebody told me, "Oh, it's a 3D printed save button." like, "Oh, wow.

Wow. That's Jenzi."

Yeah. To make everyone feel super old. [gasps]

Okay. Thank you so much. Ciao.