Adrian Brink [Anoma] x Ethereum Cypherpunk Congress

all right so today I will talk about the sovereignists in the room versus the globalists and I think this is what most people actually get wrong about the space and it will hopefully become clear to you why that is and another perspective that we should be thinking about when we think about what kind of infrastructure what kind of tooling you're actually building um so this isn't about better decentralization or better privacy or better Global da those help but they aren't enough uh this is about an ether that we should be following when building our systems and like I'll start with maybe one or two examples here which is every if you are building an application in the space and like the first thing on your website is you have very nice SDK documentation like you're probably building the wrong thing this very much most likely means that you're building a web 2 service that you expect someone to hit your server for right like if this fundamentally relies on you running a bunch of infrastructure that everyone in the room relies on whether it's custodia or non Costo doesn't actually matter but like if you're the single entry point uh you are building something that's closer to web two than to web 3 the other thing is our goal is to build Sovereign infrastructure and actually I think as a goal this is much better than saying we want to build decentralized infrastructure or we want to build private infrastructure because everything always depends on context for example there is no compliant privacy I mean this is like it's an oxymoron because you must Define under which specific real world model are you compliant you most likely cannot be both compliant with the US government the Swiss government and the Chinese government just think about this in the context of tornado Cas or privacy pools you must pick a specific block list that you are using um which means we should be building many systems that can be interoperable with each other so that everyone individually can make their own choices on whether they want to be compliant with the ofac list or I assume the Chinese government must publish a similar style list right like so it always depends on context there is no and so every time I read someone is building a compliant privacy system like cool what you most likely mean is you're building a US Government compliant privacy system which is not inherently bad it's just very specific that privacy depends on context or compliancy depends on context another example here is actually identity uh identity very much depends on context I think what most people in this room take as identity most likely is sort of national identity passports ID cards right but they represent a very specific snapshot of an identity um if you actually want to for example get a much better identity from me the right person to ask isn't the German government it's like my kindergarten or my Grandma's friends they have a much better sort of notion of who I am as a real person than a specific passport can give you right so identity really again is very context dependent um okay ah right sorry I need to point towards you yes all right the space so the goal of the space is to be Sovereign and if the clicker doesn't work because it's not on press hard no not hard pressing um anyway if you just jump like two slides forward I think because uh okay one back I want to that slide the right one um so the goal of the space is to be Sovereign and a single decentralized network is still very much centralized um and actually I'd like to start with sort of a a survey here uh who here thinks that a one world government is a good idea yes see this is always what happens so no one ever says that a one world government is a good idea but building a single centralized thing is very similar even if that thing is decentralized because single anything that is a single entity can become compromised either due to technical reasons or social reasons whatever you have so you always you never want to have a single thing of something you always want to have many similar things so that you can effectively vote with your feet and exit uh this is I think why most people don't really like the idea of a single one with government because I think in the back of our minds we always have this idea that you know if [ __ ] hits the fan where you currently live there's always other places that go to go to that aren't exactly the same um and so we should also not just design a single global decentralized system but rather many globally decentralized systems or many Global interoperable Systems that sort of work within the specific context where humans want where specific people want them to work um yeah so a single decent network is still very centralized um sort of as a quick tillon I think a bunch of people will be very unhappy with the statement but from a phys perspective I give Bitcoin and ethereum like a 0% chance to survive World War I um because most likely Communications bandwidth is going to be incredibly scarce and the first obvious Targets in any sort of large scale altercations are going to be five optics cables followed by satellites and so the remaining bandwidth that is going to be available which it's not going to be zero some Telegraph lines will still be fine um will most likely not be dedicated towards our magic internet money and getting to Global consensus on them um so from a physical perspective a single decentralized system doesn't actually solve really our problem it just means that we are now brittle and in a previous talk we actually spoke about this brittleness that you want to have resilient systems um from a social perspective they can easily be captured um normally I give this example of Bitcoin uh which I think a bunch of people will also not like but to me Bitcoin became captured because it at least the idea of it it needed to evolve over time like technology is never going to be stagnant so you want to have an asset that may have some specific characteristics but you want to keep evolving the base layer technology or the utility of what you can do with that asset um and yeah as the survey showed we shouldn't design blockchains like a single one world government because apparently no one in this room actually does like the One World Government option um yeah so a single security instance can extract brand that can harm users you always must give users the ability to leave um and here actually or you must give users a credible threat and I'll use the example of Switzerland here which is that if you ever come to switzland let me know it's a beautiful place which does some funky things such as two cantons still do public hand signal voting and it's not a secret ballot it just once a year everyone in the Canton everyone that has voting rights there meets up in their local Town Square and then they stand around in a Big Town Square and then they decide on the next on the tax rate for the next year by a public hand vote um and the best part of this is they don't count uh there's a guy on top of a tower in the middle and he estimates uh and he goes like the eyes have it or the KN have it and it gets even better because when the guy isn't certain he just gets four more people on top of the tower and then whatever those four more people say like whatever the consensus is between the five is what the vote is uh there's because and this also I mean there's like some technical reasons for why this is the case like you can always enter and leave but it's also due to having a public voting system in this case is it's actually very very difficult to corrupt it it's much easier to corrupt a secret voting system than for example this kind of voting system because you cannot obviously miscount right like everyone some people come up to me and go like well the examp what if the guy goes 90% vote Yes and 10% vote no and he still says the NOS win the guy gets chased over town like there's a credible threat here that like keeps the system stable and the nice thing is also that these referend or these votes happen once a year and you can recall your government in them so the population actually has a very credible threat to keep the system stable because they can't do vasty unpopular things because they know they will get recalled within sort of six to nine months via a public vote and on top of this you also have referendums next one um oh this actually may have worked now so also different people want want different security models and I think the best example is I keep thinking of my parents here which is my parents will never trust Global ethereum there's no amount of money or education you can do in the world to convince them that Global e theorum is more secure than their local government um this may not be true for everyone that's obvious like if you live in Argentina I totally understand why you trust Global eum more than you may trust your local government but fundament different people have different preferences with regards to their security model and this is fine like we should be supporting this kind of idea because it actually gives us in depth resilience it doesn't require that we must all globally agree on a single security model it means that different people can choose and as long as those different systems have a homogeneous architecture between them actually moving between those security models is very very easy by the way as an aside this is also I think why there is a sort of much larger Traction in the non-n Western World with regards to adoption because people generally don't trust their government as much um so like if you live in Turkey it may make a lot of sense to hold a lot of your assets on global ethereum uh if you live in Germany or Switzerland like you trust your local banks you trust your local jurisdiction your local court system right um so fundamentally different people want different security models so my parents want something like local government argentinians may want a global government the really important thing is that this sounds really good in theory but when we build this we need to really be careful about not splitting the ux um and one another good example here is phones the nice thing is I can use my phone that I bought in New York in Switzerland as well as as in Thailand um because we've sort of magically agreed up on decent Baseline standards for wireless communication which allows my phone to support to roam across all these different specific security instances right um and this is also really going to be the trick here that we want to have infrastructure we want to have um protocols that allow us to have um a homogenous application architecture that can be deployed in many different environments so that where it's very important that users don't have to sort of like point to new RPC or like download a new node software to do it but rather can Define in code I would like to get my intent settled on ethereum or tomorrow I'd like to get it settled on sanao or the next day I'd like to get it settled on global Ana uh this is sort of really the fundamentally important thing here that users and applications can easily roam this doesn't help if we have just 500 independent incompatible systems that now I have to carry around 500 phones because we couldn't agree on a global telecommunication standard uh which brings me to sort of the second part of this talk which is around scale free networks I'll just do this from now on instead great um this kind of um homogeneous application architecture in heterogeneous deployment of heterogeneous trust models is also incredibly scalable um because State should be collocated only in only if it's valuable together there's absolutely zero point in collocating uh all for example transactions happening in Bangkok with all transactions happening in Zurich um if if they do not touch the same piece of global State why are we putting them into the same synchronous Global consensus lock it we're just paying now to like keep a bunch of State in global consensus we taking a bunch of latency hits so State should only be collocated together when it's valuable um and generally this means when it's culturally close as in and I say culturally close here because the World of Warcraft Community has a bunch of state that is valuable next to each other um the ethereum commun we have a bunch of state that's valuable next to each other someone doing random payments is not valuable next to World of Warcraft or etherum state this is like you want to have the ability to have very fine grain control over where State lives and on what security model State lives uh because it's going to be much cheaper incidentally it's also going to be much faster um by collocating only the state that you actually need for example um if we only in this room want to trade the single fastest and most optimized way to do this is to do local area networking um salana will never beat this unless they can violate the laws of physics um like if we want to get the if you have latency critical applications it makes sense that you are directly connected to each other and this really comes down to physical proxim proximity because Global connectivity is not all that fast it's maybe something like probably 500 milliseconds or so if you're fast um so you should be able to only opt in to alter cating State when you actually want to pay for this for example if you want to execute a block trade between Zurich and London for example between two parts of state that actually like you want to have an atomic update on but that live on two separate systems then you should just be able to opt in to update uh to like collocate these two pieces of State you shouldn't have to force everyone to be collocated with each other and then users in application can actually start roaming between um instances here the interesting thing is what happens when you start thinking of users as um their own State Controllers so state is always created by someone for example Circle as in usdc um they are a one out of one blockchain and really the easiest mental model to think of circle is a one out of one blockchain that is backed by a single key that sort of gives updates to everyone signed by their key uh and they create usdc right ethereum creates eth um no one else can inflate the supply of eth by the theorum um and so you always then start thinking well if I create my own token so if I have admin power over some sort of asset that I issue it would make a lot more sense if it's running on my one out of one blockchain that I'm that I'm delegating this to some other external system where I still maintain admin control um y u yep so every Community can collocate their personal State um and as a result single instances only need to handle sort of the maximum State SL user throughput of that Community uh you can then turn this an on demand charted system if you're actually interested in the consensus research behind this uh there is a bunch of papers around Typhon um and I really mean Typhon because I thought that's the way you pronounce typhoon um so it's called Typhon now and also pron yeah it's called It's pronounced tyen um and yeah at the example of the block trade between two stock exchanges uh yeah this is actually also the other reason why this is super scalable is because when you think of the routing problem of if every instance if every chain is the control of its own state it makes a lot more sense to Route via the chains that create the state so for example if I would like to Route if I like to move some eth from optimism to arbitrum I should obviously route this via ethereum if I route this via a third party it makes absolutely no sense so as a result like if you want to route usdc from chain a to chain B you should always route this via Circle directly because it's like the least trusted um way to do this because you trust Circle anyway to do it uh it's also going to be the cheapest because Circle needs to update its state anyway um and the super nice part of this it gets rid of one of the really one of the we abstractions we have in the space which is between users and chains because right now there's a user and the user must connect to must run some full node or must connect to some other external full node and issue RPC requests against this it is much clearer if we think of users as one out one blockchains that have limited State uh they only have state that is relevant to that user all of a sudden we building exactly the same networking Stacks the same interoperability stacks that allows users to move some assets between them and ethereum and them ethereum and optimism it's like we just get this very nice homogeneous architecture where we don't have to jump these weird abstraction barriers which we currently have to do yeah and I gave the example of asset routing yeah the last part is the goal of the space is to build World War II resilient infrastructure and I'm using this very extreme example of World War I resilience because I think it is not because we should be hoping for this this would be very messy and we should all pray that this doesn't happen but it gives us the right frame of mind to actually design Brazilian infrastructure because it's a very specific secur um sort of like threat model um so fundamentally we need to provide a replacement upgrade for existing coordination infrastructure um we have historically lived pretty much the internet has lived in this very happy pakan Gala and none of this infrastructure was designed um living in Switzerland I think I can confidently say that the Swiss Financial system will not survive a dedicated onall by any sort of nation state actor because it was never designed to be actually resilient because when the internet came about it was sort of like oh this is this very nice thing everyone is very friendly with each other um and we didn't really design this in mind with having large SC like having critical infrastructure on it um on day one of World War I Global connectivity is immediately gone I used I gave the reason sort of why earlier on and this fundamentally means that things like visa and MasterCard will just stop working because they to a large extent rely on global consensus I mean not Global decentralized consensus but some amount of servers being accessible um in their sort of in the US for example this is also the reason why I find the moves that Sweden and Norway are making with regards to uh cashless societies quite Brave um if you're going for a Casher Society you better [ __ ] be sure that your financial infrastructure is actually resilient in sort of the future because the future is going to be more unstable than the present and I think like we're heading in this multi into this multi poar World Order where things will just get messy and more complicated and we need to be designing our infrastructure for this case um as a result we also need privacy but here I would really urge people to think of privacy not as something where you are fighting the government or it's privacy for National Defense or privacy for Community defense I think a lot of the reasons why people think of privacy should be against the government is because they secretly don't or openly don't like the government but like this is not really like if you don't like your local jurisdiction move vote with your feet right like or try to change it locally on the ground but really think of of this that Switzerland the country has a vested interest in its financial data not being leaked to North Korea um like I think the quickest way to shut down Global ethereum tomorrow is to move all the US Financial system to Global ethereum today and like in a day or two the NSA will go that's probably a bad idea why did we make our our entire Financial system readable by the world including every foreign uh Power so really think of privacy not as something that like think of privacy as something that every Community will want because in order to have stable communities you need stable Financial privacy like my parents will never want to adopt a system that leaks all the data either to their neighbor or their all the North Koreans really um but I think privacy for National Defense is a much more sort of positive framing of the problem than um like you want to hide your financial information from the tax authorities um and the nice thing is in the best Cas like when I think of World War I resilient infrastructure it's it's a little bit about in the best case we never need this infrastructure in the best case we always maintain Global connectivity but in the worst case we want to have infrastructure that is capable of running as local as required but as Global as possible um and this really means for example if something happened to all connectivity outside of this room it' still be nice to be able to run coordination infrastr structure that is local to this room rather on like the computer hardware that we have in this room and this is technically completely feasible if we design our systems and architectures around this um yes so fundamentally coordination protocols IE blockchains can interoperate as much as the physical Network allows if you have Global connectivity you can do Global connectivity stuff and you can like rely on global consensus if you only have local connectivity you can only do local things but you want to allow the systems to gracefully degrade and upgrade again so that if you go from Global connectivity to only local connectivity you can still make State updates over your local state and then as you recover you can reach out into the global system again um yeah and scale free networks are all about being as local as required or AS Global as possible um and I'll use the second example of the Swiss here mostly because the Swiss are paranoid and the swis have been paranoid way longer than our space has been around so I think to some extent we should be learning from them a little bit um and once in a blue moon when I moved to zuk I was like oh I should get involved in the local community and I looked up the ham radio club of zuk and so you go to the website of the local ham radio club and then you click around the website a little bit and then you come across the charter given to them by the Canton of zuk which effectively states that the local ham radio club is responsible to get emergency backup coms online within 6 hours of a national of a national disaster and the entire idea is that the local ham radio club has volunteers that have cars with backup generators and antennas and when [ __ ] hits the fan and you lose sort of connectivity they will drive them up to the mountain peaks around the Canton three minutes great I'm almost done um and restore backup coms for emergency services and this is the kind of thing we should be thinking about like if we actually want to have resilient infrastructure we need to be thinking like how do we actually build these resilient systems and I know this talk isn't for those in the audience said like mean coin trading and like this is much more sort of like to some extent depressing but also I think this is where the real value from the space comes from like the real value is not the mecoin casino the real value is the resilient infrastructure that we can build along the way um yeah oh this is maybe the really last important point if you're in this room you're most likely in charge I.E you in charge to provide local infrastructure to your local community um and this starts with very basic things such as please run your own eorum notes if you're in this room into your Technical and you're not running your own ethereum noes like we're doing something wrong in the space like everyone should be running their own ethereum noes everyone should be running their own unop front ends there's really like if you are riant on infura being online on unisoft dorg not being compromised this is a bad setup and we like should be practicing the things we preach um and yeah for your local community you are most likely the guy in charge um because like my parents will not figure out how to install a new application on their phone or how to run in ethereum node but someone in their Community or will right like it's not that everyone needs to be completely self-sufficient generally speaking societies don't work this way there's like some division of labor and you want to allow like we are in this room are the people that have to rack up a bunch of hacky Hardware to make local consensus work again yeah so quick teldr um single decentralized networks are still centralized a single Global Security model is [ __ ] terrifying and please not let's not sort of back to our way to the One World Government um because we don't like it for many valid reasons um different users have fundamentally different preferences for their own security um and we need many independent security models in order to allow people to have a credible threat of an exit um fundamentally when building these systems we should be thinking about how do we build scale free networks I.E things that scale um very easily because we aren't design designing bottlenecks into them and like for example in the asset routing example if there would be a single chain to Route all assets that' be a very bad idea because that single chain is now scaling bottleneck so you want to really design these kind of scale-free protocols and architectures in order to um get there U yeah you should have applications that are capable of roaming privacy really matters and please run your notes honestly if you only remember like one thing from this talk go home and set up your own local infrastructure uh it's not hard a single Raspberry Pi is all you need like single Raspberry Pi and a 4 terab USB drive and you're good to go to run like all infrastructure you will probably need in the next year um and if you don't know how there are many people today and also through the rest of the week that will gladly teach you how to do it um and I'm saying this from a very personal thing like the way I got involved in ethereum in 2016 was at meetups people would teach you how to install your own ethereum not and run your own ethereum not before you could make a single transaction like the first thing you did was like install ethereum run the note sync the network and only after you manag to sync the network you could start actually using ethereum it was it was great um I think uh yeah we should also learn a little bit from the internet uh the internet was actually designed quite well and yeah the goal of the space is to be world with 3 resilient infrastructure thank you very much