Juan Benet 🛸 Preventing Digital Totalitarianism

[Music] I'm going to talk about today and to kind of warm up the day and get us going um is I'm going to talk about how to prevent digital totalitarian States uh I'm going to talk about how these might rise what the current Powers might be and what we should be solving for uh it is not a given that we're going to land in these kinds of things it would be great to avoid it but it is a definite possibility and a lot of Technologies are empowering these kinds of structures so um thinking about this kind of thing uh I like always going back to 1984 because it has a very good mimetic power in terms of um in people's minds what uh what a society with hu huge amounts of power and control over Society um how that Society behaves if you go back and read the book it is super tame like technology plays almost no role it is primarily social and so when you reason about um 1984 like the Technologies in at play are maybe like hey the the TV in your room is owned and can look at you and maybe there's some amount of surveillance in CCTV and it's just like the prediction relative to where we are now or not necess prediction but that model is just so so tame with what's possible today so if we were to write to you know George Orwell's uh 2024 hour 2024 um what would that look like and so I'm going to take us through that um in in a bit um just to give you to to have a little bit of a of a claim to why I know what I'm talking about um I've spent the better most of the last 10 years uh working on a set of Technologies to prevent these cancer systems uh these kinds of problems um ipfs and lipop which are um tools uh designed for a much better um internet infrastructure that can give us a certain level of um safety and human rights uh so human rights is it's a huge focus of of our work um and we try to create structures that protect civil rights globally or give people the tools to protect civil rights globally um it's not given that distributed Technologies actually do that and help you it just empowers the people locally to make better decisions and it gives an asymmetric Advantage um I think everyone here working on on web 3 uh initially got inspired by that kind of vision initially got inspired by by bringing in um those kinds of Rights and I think that web 3 has an enormous amount of potential in uh achieving this sort of thing um some of the tech that we've built uh was we were so no strangers to censorship that we were involved in many many events um our Technologies got involved in many many events relating to censorship so uh what are modern orwan States what uh could be built uh then we'll talk about how to prevent them then after that we'll talk about building Global secure Communications I probably don't have to convince most people here that totalitarianism is bad uh but to put things in perspective um you know think of completely locking down some Society um and preventing any kind of U shift out of totalitarianism um but also just kind of when you weigh it by the the horrible deaths involved you just kind of end up in a in in a just terrible terrible Outlook um now like what's new in 2024 uh so first off telecommunications globally has have d drastically uh improved since 1984 uh we have now the internet for most of our interactions we have about you know 7 billion smartphones deployed we have CCTV basically in most cities uh in certainly most places of of Transit um any kind of hubs for for moving around our full CV we have satellite imagery uh globally you have very good resolution in those um pretty quick updates we have a fully digital economy that could enforce policies and rules which you know if you um suddenly are disliked by your local local uh system and prevented from moving around uh taking some public transportation buying food water whatever this could get um a little hairy uh and of course uh AI uh is going to make all of this kind of stuff way way way harder uh think of levels of you know massive all of that massive surveillance uh needs to be analyzed by somebody humans are not exactly the best at this uh computers are extremely good at it so think of all that that object recognition and stuff that we looked at now deploy it at massive scale uh being able to then guide behavior and policy you can build and deploy automated systems of control uh you can have behavioral ml models um we're going to talk about a little bit more about AI because um since I give the first version of this talk earlier this year uh there's been a set of improvements worth including um you know we I mentioned surveillance but you know just to put things into perspective um we have lots of evidence that many organizations uh around the world and monitoring all traffic um metadata and the content uh phones GPS positions Audio Video Communications data Health Data all that kind of stuff um if you're like me and you really like tracking your Biometrics you have one of these devices and you take all of your data and you give it to the cloud in the clear and so you know several probably several corporations that right now have a full biometric model of me uh and can predict uh how I move and so on uh other corporations have full models of what I say online and can predict what I'm likely to say um you know it's not not a good not a good position not a good outlook uh then you know beyond that you we have all of our Communications and social media like just all of this is mediated through through currently still corporations and subject to a lot of local policies uh we've had uh evidence of a bunch of propaganda problems uh control of news uh Publications and so on disinformation um and you know if you think about modern Computing and the ability to generate content today um and kind of coordinate and if you have a system that is uh producing content optimizing for some goal and is measuring uh the reactions of people uh that can run off in a pretty terrible way you should uh pay attention to what happens in the next us election that'll be a really interesting death bed for uh a lot of this kind of stuff hopefully nothing uh really bad happens hopefully a lot of the hopefully good will prevail and a lot of this won't won't come to pass uh but it'll be an interesting place because people are certainly going to try uh then we have things like um you know if you if you have all of these Computing systems and you can uh control a lot of robots and so on um in in a in a in computers uh you can also do the same thing with humans you can uh measure the behavior you can create some incentives um and you can couple the surveillance uh system with a control system uh we're starting to see the beginnings of this kind of thing being deployed we have the social social Credit Systems in a few places in the world these are like super tame right now there haven't been really weaponized um and one of the things that I think is crucial to enabling this kind of control to happen is that these things get implemented and then just make everything better if these systems come in contact with a country they make everything better for years everyone learns to love these things they're like oh man like the world just got so much better after the social Credit Systems more and more and more societies will just grant control to these things over time a decade or a couple decades in transition the administration's transition somebody starts rewriting the rules in the algorithms and boom now you have a a really bad situation one of the key things here is most administrations uh in the world today tend to want to protect people tend to be pretty well intention we we have have had a relative relative level of Peace compared to other parts of History um and so usually those administrations are trying to be defensive um and implementing extremely bad systems that are going to be taken advantage of by successors like this is classic history like just go back and L about read about the histories of the the worst kind of um um police systems got usually implemented with with very good intentions uh then we have things like you know automated enforcement of of of situations right so not just incentives we have robot storms coming right like this this stuff is gonna get nasty uh but this is about censorship so we won't dive too deep into this promised AI uh AI Edition uh so uh a lot of stuff has has changed uh we have again massive surveillance and surveillance and data analysis we can generate an enormous amount of media um you know llms are just tremendous L powerful just try playing around with them it's amazing um you can generate and summarize all kinds of content you can uh we'll have we have enough computing power to just observe and write for All Humans right so the computing power available to major corporations and to nation states is sufficient to just generate to to observe all streams of communication coming out of a human and write back to the human that's insane how do you even like come you know fight that you we have to bake in um uh protective systems by the way this is like first part of the talk we're going to be very optimistic towards theend don't worry um uh and then you know think of like new things like we'll have New Economic structures new control structures we we have like preventative control models for like are you likely to dislike the state in a year uh should we like prevent you you know uh cut that early uh and this isn't even touching AGI right like that's way out of scope for the stock in that you know it's coming predictive people predict is going to come in like the next two to five years so you know very out of scope for today um but you know if you want to wrap your head around this think of anything a human can do at a clock clock speed a thousand times faster fully connected to the internet like that's you know think of that that's sufficient of course these systems are going to get better so um problems so how how could these so so we' talked about the capabilities how could the totalitarian systems rise rise usually um there are a few different ways but usually they tend to be um they tend to rise in the name of Public Safety whether it's from dictatorships after some upheaval or democracies where something bad has happened Happ and so groups uh tend to uh vote in all kinds of safety measures um the Patriot Act in the US is a very good example of this uh eventually even though those administrations could be super well-intentioned and trying to do what's best for everyone by creating infrastructure with that level of power eventually someone bad is going to look at that level of power plot a um plot how to take over do it and then Implement a a really terrible system you just have to pick up a history book to see lots of examples now practically um what this means is implementing pieces of these systems demonstrate that they're really good for everybody get everyone to love them and want them to spread and increase that distribution over time you keep up ratcheting power and control and eventually someone will come along and take over so great we've talked about like the problem how do we prevent these terrible things um so there's a lot of different ways uh you you need a full combination of approaches primarily you need a super well educated public that understand understands the potential um that has reliable media this is where censorship resistance is really critical uh you need you know schools and learning and you need humans to be very well aware of what's possible here you need extremely good policies in government you need balanced government structures very strong institutions that can prevent this kind of takeover very strong institutions that are like red teaming problems and trying to um nip things in the butt before before they become an issue we also need extremely strong um technological Powers we need extremely secure Communications this means again censorship resistance to enable all of us to coordinate um but also privacy uh reader writer privacy being able to coordinate requires the ability to coordinate um uh in private so that you can't be taken out by some wouldbe totalitarian controller uh and this usually gets forgotten the E economic Loop here is extremely crucial you need a very strong economy um to um prevent like these massive upheavals that tend to you know give in stability to governments but more importantly from my perspective you need a very strong economy to stay ahead of the curve in the technological um stream to be able to implement all these Technologies with human rights at the Forefront um instead of like lagging behind others and then letting others kind of implement their control structures on you um I won't go super deep into this because this a little bit of out of scope for the for the uh for today uh but you know the world has hundreds of democracies there's very strong institutions most of them really want to prevent this most of them really really are aware of these kind of problems and want to prevent them um but there is this big debate in all of them about between security and privacy they are very real short-term practical problems with security and that usually gets used by the would be um would be implementers of these kinds of appara so um the world has used thankfully the Cambridge analytica problem gave us a very good example to use now to talk about these problems to implement safeguards so this is you know no matter what government you're you're in and so on bring this kind of thing up and um yeah fight for this kind of stuff now I want to mention that privacy is not a great brand it tends to be dismissed as optional it does not convey the real problem we we as a community of Builders you know going back to the 50s and so on um missed up here it's not about privacy it's not about keeping what you're doing Secret from other people like that's that's not at all what what is going on we're trying to establish security um secure Communications is what we're talking about being able to coordinate securely as a people um that's what we're orienting for and so really when you when you whenever you kind of think about privacy and so on really strongly mention security because that's at the end of the day it's it's about two forms of security short short-term Security in terms of um current problems or long-term Security in terms of um how your civilization is able to prevent these kinds of problems um you know there's a lot of good progress to highlight you know there there were the original crypto Wars uh strong encryption standards they always come up for you know people always try to get rid of them but we we have so far been prettyy successful um there are really good proposals against AI biased um systems um we need to like go hard on this um stay ahead of the curve here um and I think in general we we just have lots of lots of groups fighting super hard to to prevent these kind of problems um huge Kudos too you know tons of people around the world uh that have worked super hard to to enable this uh great so this is technology conference so let's get back back to that um the I'll dive very deep into Communications in a moment I just think about it in terms of like we need secure Communications talking and writing to each other in order to properly coordinate so that we can take significant actions and prevent these kinds of problems and so the basic Foundation that I want everybody here to walk away with is that we need Global secure Communications censorship resistance uh reader writer privacy we anybody any human in the world should have the right to communicate with any other human in the world securely and privately and in groups so U again I talked about the economy um I won't dive super deep into that but just um the the economy is what lets you coordinate large groups of people it lets you build the technologies that you need uh and it lets you take actions uh cool so let's dive into uh secure Communications I'm going to give like a a view into the internet infrastructure lots of applications and hardware and so on and I'll kind of hint at a road map of what we need to do so at the very bottom so I this is not typical in in the blockchain space we tend to operate kind of like on top of the internet and the internet is this big black box that we don't think about oh yeah just send packets to each other right it'll just work out absolutely not sensorship sensorship happens at a lower layer we need to improve the entire IP stack we need to bring in the blockchains but syn it way SN it way all the way down into the basic building blocks of the internet we need um uh proper uh communication protocols that are able to have um uh censorship resistance directly in the in the uh routers in the in how your computer connects to the internet and so on we need to make this a human right like it's not going to be easy to implement this globally we need the UN and other groups um behind uh uh these kinds of um policies we need things like uh you know secure uh systems like T ITP and like the new mixet like Nim and so so on I want to call those out as like really good efforts uh but we need something that's going to scale to the billions and you know tens of billions or hundreds of billions of devices that we're going to have um a lot of these systems are really good for small use cases and they're really good to like um you know give us something to work with and scale but what we need to be targeting is something that's going to work for tens of billions to hundreds of billions of devices uh communicating a lot so when you think about protocols um really Target that because if you're going to spend three to five years designing a protocol and and implementing it and five years in you're like hey great like you got a 100 million things to communicate privately awesome that's a really great progress but you know like the there's a huge gap there um then you know kind of if we if we manage to you know kind of getting sensorship resistance and privacy at this layer is extremely difficult it's going to be hard to start here we need to make progress on this but but it's a an area that is difficult to to get to um one thing I would recommend folks here is create large scale um incentivized routing systems so use blockchains to build large scale incentivized routing systems in the clear at first and get them to work and get them to scale um then you can you can bring in the the sensor resistance and the privacy reader writer privacy um later but just showing that you can build a a massive scale routing system that works globally internet wide would be a great plus um you know prop uh and then you know sell sell networks and and other kinds of routing um and satellite and and whatnot would be really useful it be great to have like a like a decentralized star link type of thing that would be uh super valuable that would enable us to you know really communicate with a lot of people around the world cool so um DNS is one one really strong aillis heal and some of the things that I mentioned before uh this was usually a lot of States tend to attack DNS because that's if you can't find the content and you can't don't know where to go tough luck you even though you could theoretically communicate to the right computer um you can't find it so DNS is like this place aillis heal that you can like cut in terms of um how people find things uh this is where things like blockchain naming are really useful things like ens um and so on are extremely valuable in terms of ratcheting this up but most of those systems still use the traditional routing infrastructure we need the names naming system especially for really sensitive names to be able to be spread out and distributed to lots of devices so you can load things locally uh without having to depend on some some registry later that could be blocked um we need to radically improve the internet um Sor the ca system so um in blockchains we thankfully use much better cryptographic systems but most of the internet and most of the places where we get our software are still using the ca system so great you have a great blockchain awesome Kudos uh if you can't get the right software and the software that you can get is owned by um by somebody attacking the ca system uh tough luck you you now have a vulnerability and like um you won't be able to to do what you want it anyway so we really need to take the magical crypto systems that we have and rewrite and and improve how our browsers get content in the first place um we need to we need a way better um web that that has censorship resistance baked in um this is stuff that a lot of us have been working on and we've made enormous amount of progress still a ton to go we need the stuff to work on every native browser um and we need it to work at scale for all kinds of applications um right now work really well for static content um and so on I mean lightly Dynamic content but it doesn't yet scale to things something like a like a social network that though you could do it in practice um uh it doesn't work well enough to to um to really spread um that's a major area of focus for for these projects uh we really need browsers to be be very strong privacy oriented we need hooks for routing systems for naming systems for certificate author authorities for other loading systems that load other content Sy systems the load applications all of that kind of stuff the browsers and the operating systems are these key entry points where if you can hook in there um then you can distribute much better much better infrastructure um app stores are a huge problem right now because um you know Android is a little bit better than than Apple but these are basically super locked locked in environments where you can't uh distribute whatever you want onto these things uh they're super controlled in terms of um geopolitical structures and so we need some way of having um the the control over the mobile phones uh networking systems directly from decentralized App Store type of environments not sure if Android enables this right now it might because Android has like a much healthier open source orientation uh but I know certainly Apple um does not let you install anything you want uh you can only install what Apple wants and you know I'm going use Apple devices all the time but this is a huge problem and they should improve they should know better um then you know you want to think about like the different modes of communication and for each of these modes make sure that you have a really good awesome product that everyone loves that is secure uh by default that has censorship resistance that has reader writer privacy One to One Communications one to many many to one many to many kinds of things uh you can imagine although you know think of any me application that you use for for communicating with other people um telegram signal um email text messages uh forums whatever you should have a an alternative uh that works really well that is beautiful and uh wonderful to use um that censorship persistance and privacy rear writer privacy Bon and and kind of um even I would encourage starting with very simple thing so like the fact that we have signal and being able to do one to one or small group Communications already is a huge plus the Mana data is still observable by attackers you can see you can correlate uh who's messaging who and so on if you're observing the link links that needs to get way better um and then you know on top of all this you know think of like much more complex applications once you have economies and organ uh organizations and what not you want to enable those things to write on top of this these kind of systems uh eventually it would be great to get the secure Communications Hardware um there's the problem here is that there's very few vendors that can produce the really high quality stuff there this means an enormous amount of lock in uh the good news there is that there's a lot of like there's a strong orientation towards democracies and civil rights which is awesome um we need to get we need to work towards getting more vendors um and uh getting these systems to open up so that you can run whatever um whatever you want in them uh and I'll give a plug here for research and development in general so um the what one of the things I love about the crypto Community is that we rle the amazing cryptography papers and we're like oh my God like this is amazing why don't we have this deployed and you know old cryptographers were like when you deploy new cryptography it's bound to break and never works and that led to this massive slowdown in in uh in cryptography so new things weren't getting deployed and developed we're like hey let's take these magical tools deploy them at scale um and if they break we'll fix them um and let's use them to generate an enormous amount of wealth that we then will Lube back into the R&D cycle to make even more cryptography and even more systems and that has been amazing so from a perspective of like um what is the thing that has happened in the last 15 20 years that has been the most promising for secure Communications it's the rice of the crypto um crypto Community like the cryptocurrency community is is feeding the entire cryptography community and accelerating the research into Z proofs into fully homomorphic encryption into just secure Communications in general distributive systems and so on so we need to like double down on this uh it'd be great if blockchains in general started devoting a huge fraction of their you know significant fraction of their uh wealth directly to this kind of R&D like funnel like emission of tokens directly into funding R&D of this kind of stuff you you'll hear me talk a lot about that over the coming next coming years um there's really great new tech Frontiers a lot of the speakers here are going to talk about this stuff Z knowledge proofs is all the r right now awesome um not good enough for certain kinds of things I think things like fully homic encryption are going to be critical uh these are going to become much more um uh widely deployed and and important in the next 5 years uh so keep an eye out for that there's a lot of work to do there um I think that these things can be used for routing for all Communications for all messages for all um um yeah for all kind of like normal human Communications so it sounds crazy but when you think about the number of thoughts and messages that a human can output um you know within 10 20 years we'll have enough computing power to run all of that in in Secure Communications fully with like perfect cryptography so don't worry too much about like things being way too slow that's that's just wrong um just wait wait a few years uh it'll get way way way better um and like I mentioned the Innovation cycle we need um successful uh projects um to to scale to generate a lot of wealth so that we can feed that back into into R&D um and I I just want to kind of um end with a huge congratulations to the entire ecosystem because overall we've made an enormous amount of progress in the last few decades um especially from my perspective in the last decade um we did that on the shoulders of lots of giants that you know fought the crypto Wars gave us basic cryptography highlighted all these problems um but I want to kind of remind everyone that we need both production grade infrastructure and software um for lots of different use cases and you want to find things that are going to be really important and you want to make sure that you have a beautiful wonderful product that everyone wants to use a lot of us use and love telegram because it's just so nice and really useful um but I don't I don't think the whole thing is open source and without knowing that the whole thing is open source and secure and so on with like really good cryptography I can't trust it um and instead you know signal is like a little bit of a ratchet better but not dramatically better still because you can't um I think you can't run all of your own instances I think some of the um like again it still relies on the app stores which kind of suck um and it's still I mean you have to go through the app stores but you should offer alternative ways of communicating and the metadata is kind of in the clear um but but again super great that we have those things and um all of these things need to get better uh and we have a really good Innovation cycle it's really special that the cryptocurrency community is has this entire thing going we need to scale it uh but just as you head into the conversations today know that this is there's a this is a long road it's a long fight here uh we have a long long way to go to we're fully orwellian proof and digital totalitarianism proof um we've had an enormous amount of successful progress we need to double down on it and scale it we're in a dramatically better place today than we were 5 years ago or 10 years ago we have you know magical internet money we have much more secure names we have much more secure ways of communicating our applications and our web pages and so on we have pretty good sensorship persistence relative to where we were 10 years ago um but we have so much more to do uh and so I really hope that the people in the room uh are going to uh sign up to um come build these human right protecting Technologies and help us all U live in a much much safer and much much better future uh thank you very [Applause] [Music] much