Coordination Is All You Need

Transcript

so my goal here is actually to talk a little bit about coordination which is adjacent to the public goods conversation um the goal of sort of talking about coordination is actually that it turns out all coordination or all public kids problems rather can be sort of simplified into forms of coordination problems across different jurisdictions whether those are State jurisdictions corporate sort of boundaries just any really sort of Institution let me make sure this works and why should we even care about this I think people I think by that by virtue of being this room understand the idea that it's all coordination people have heard that phrase before but we've done a lot of really cool things as a society

that we don't really think about very deeply and these are all the way from doing aqueducts in 2000 BC to doing wood block printing which actually preceded uh the printing press by about 700 years to doing things like the internet in the 1980s and these are things that we care about because we're able to actually build new forms of abstraction around our Civilization by creating them and then sort of achieve our broader goals by using them in their day-to-day lives and why this matters so deeply is it's actually all about coordination when we think about building these things and we're at a stage of development as a civilization where ultimately we're moving from sort of Kingdoms in you know the last 500 years to states in the last

sort of 200 years to networks in the sort of Modern Age and the major problem is that we don't really have better ways to coordinate these networks in the way that we need to in order to actually solve these kinds of public kids problems or other broader classes of or types of coordination problems so we're in this existential sort of period where we have threats like you know nuclear war climate change artificial intelligence and we need to sort of figure out what are the ways in which we coordinate each other around these sort of Concepts and this is particularly hard in a networked age where we have many more people than we've had before we have you know billions of people we're used to operating at this dunbar's number limit of 150 people where we can start to sort of know our local tribe or community and

as people uh sort of grow into these larger communities we have you know pretty much the Leviathan in this model of the uh the state as pretty much the last Bastion of of how we organize in the Modern Age and we're only just coming to terms with what sort of broader protocols might look like and these are protocols that started with things like speech and language have evolved into things like the internet itself and then have further evolved in the transfer of value in terms of things like gas per protocol like ethereum itself in terms of allowing us to actually transfer uh really tangible like value across the sort of states and different networks that we're part of so coordination is important we've built lots of cool things using coordination

we can use protocols in the modern sort of networked age to coordinate and we can do this at a way that scales hopefully globally rather than in previous areas where we've sort of focused on you know scaling locally and there's nothing wrong with local scaling it's just the problems we're facing now pretty much require us to have better models of coordination because if we don't all agree on the state of the world that we're in then we can't solve something like AI alignment and this is actually a pretty hard problem it turns out there's an XKCD comic for pretty much everything and this is no exception and essentially there are sort of these problems where we have great protocols but no one really agrees on them we obviously have

very different languages we have different uh sort of protocols around how we coordinate in terms of social norms and rules even when it comes to types of Institutions this is from the summer protocol sort of draft paper we have things like marriage which are different across the world we have things like masking protocols or even birthday parties which are different across the world these are all things that we can start to agree on more more carefully and more precisely as we start to basically think about coordination in more of a protocolized sense not that all these things by the way have to be protocols but the important piece is just how we actually think about the problems that we're solving in this sort of frame of reference of networks versus in the frame of reference of local communities

and so to go a bit deeper into the particular problem that I'm interested in when it comes to coordination I think aligning on alignment is going to be really important here there's a lot of conversations around AI safety uh there were a lot of those in the past weeks on the left here on my side there is a uh sort of great chart of all of the different approaches to AI safety you can see that there are many of them and on the right here on my side you've got the probability of Doom across multiple people and essentially my view is that if we're going to solve these sort of large-scale coordination problems we all kind of have to align on where we are in these dots before we can even get to the point of coming to Solutions so we we need to actually like find ways to

coordinate around the problems and this is again something that's important for public goods but also important for General civilizational uh success and longevity and the way that we come to this alignment I think fundamentally is through social norms social norms are still a pretty critical part of even the protocols that I've already mentioned but the way we come to these Norms is very opaque and very legible and I think what we need to sort of figure out and something that uh folks like Metallica done a really good job actually articulating better than I have is what legitimacy actually is how we get people to bootstrap legitimacy and how then we get people to agree on the state of the

world such that we can solve particular sets of problems and agree on those problems first then agree on the solutions later so an incomplete survey of some of what legitimacy is just from previous historical thinkers Max Weber was a particularly popular person for this sort of theory of legitimacy and of social norms and institutions he sort of defined this as you know behaviors and habits that helps out set out boundary points or bounds on acceptable interactions between individuals According to some already existing sets of norms and rules again you can see how this creates sort of a bootstrapping problem economically we can consider it as sort of a force that constructs shelling fences or ways that people can without

communication without even speech come to agreement on a shared state of the world and so for example if you've ever seen people having like a picnic at a public park the public park is a kind of shelling point for people having picnics you know that if someone's having a picnic they're probably doing it there and you can sort of know if someone tells you that they're going to have a picnic if there's really one park in the city that they're probably at that Park this is something that we can use to generate sort of pre-commitments and we can sort of get this bootstrapping done through slowly building up these types of interactions over time and then thirdly systemically we sort of think of legitimacy as a process or possibly even protocol itself that encourages these higher orders of acceptance to be built across different sets of rules and by doing that we can

sort of Advance Civilization by extending the number of important operations which we can perform without really thinking of them this is actually also from the summer protocols draft paper which I encourage everyone to check out so that's one sort of set of definitions but I actually think metallic does a good job of diving into these in more detail in uh sort of six main examples and I actually think these are in order for me of legitimacy of the actual model itself Brute Force being the least legitimate form of legitimacy that makes any sense essentially you're compelled to comply with something and so therefore it's legitimate this is true uh very historically in Kingdoms in which the divine right of kings gives anyone the ability to you know uh or

requires people rather and gives the king the ability to uh let anyone or Force anyone to do what they want and otherwise you're kind of stuck because the like the Divine rate of Kings is is literally from some God or gods and therefore um you are it is by definition legitimate you're forced to comply the action is legitimate you have something like continuity which kind of extends this into well maybe you had some kind of coercion at a start or some kind of other model that created the legitimacy but then it continued to be the case that this sort of event kept happening and so now it's legitimate at like the next t plus one time because it was legitimate before also a very weak

form in the sense that you know if things just continue long enough they become Lindy and then things that become Lindy become more and more legitimate and there's a bootstrapping problem here because you have no idea what actually is truly uh legitimate at you know t0 you have processed legitimacy which is getting slightly better here you've got like the idea that processes can tell you whether or not something has been done fairly we'll get to that in a second but if something is done sort of in this way where people agree that the process is legitimate then you can have the outputs be legitimate uh you've got performance where essentially if the outcome itself is just good for society then naturally everything is works out for the citizens of that society and so everyone's pretty

happy there for everyone to use the thing is legitimate and now you start to get into what we actually I think consider in this room and probably generally in the web three spaces the two forms of legitimacy that we've thought a lot about when it comes to the protocols we create and the ways that we're thinking about for example solving some of these existential problems I mentioned earlier the first being having people involved in the process participating in the process this is obviously the premise of democracy it's the premise of a lot of political systems we've created and it's a very important part of people feeling like their actions matter and that their voices are heard there's a great sort of set of writing on voice and exit that I would encourage people to check out on this as well

but now we get to the most complicated and the one that I think we should really start to think about in a more precise way and this is where the protocol conversation comes back what do we mean by sort of fairness and whether an action satisfies an intuitive notion of brightness in the first place and and this is an important problem because we think of this as a very abstract moral question a lot of this seems maybe abstract people in the audience but these are practical questions we should actually all be asking as we create systems like for example new protocols or new D5 you know nft Dao whatever we want to say protocols or or systems that are going to potentially impact people at a global level and we should in particular be thinking about who is it

fair to in the context of a global system because there's lots of people who may be excluded from that process who may not be able to actually access the systems who may find them too expensive who may find them too challenging to use there may be people who you know we can use public goods sort of bootstrapping to include in the process us more actively and we should be thinking about you know what this intuitive notion of fairness is for us as a community so that we can start to solve this problem in the context of the AI example I gave this is actually very challenging because you don't actually have alignment on what a kind of uh what the problem is what the risk is and therefore what is fair in terms of

inclusion and participation but also in terms of the risk trade-off that we're taking with different approaches so I want to give two kind of examples that I personally think are interesting models that people should be considering when creating their projects practically speaking and these are models that essentially will give a little bit more context on what we're doing with Git coin a little bit of what we're doing with a lot of the brighter public goods moving in web3 and the first is sort of a framework around internet native Commons in the context of solidarity economies and this is something that folks like Primavera have talked a lot with this idea of coordinations which itself is sort of just upon in portmanteau where you have solidarity between a group and a sort of mutual Aid

between that group you have some sustainability process for kind of continued longevity and then you have sort of Independence or sovereignty of each person voluntary participation in the network that allows people to feel a sense of agency over their participation and what you get when you do this is a much more polycentric functionally polycentric Network meaning that you have actually like relationships between and like directional relationships ideally bi-directional between each of the nodes rather than just maybe polycentric communities that don't interact with each other and if you kind of think back to the slide I had way before on the local to Global problem what you see is that we have a lot of solutions that kind of get us to the second little box here we have lots of

these different states we have lots of these different sort of companies that may have their own rules that may have their own processes but we don't actually have a way to get those individual rules and processes to an actual large-scale polycentric and like aligned version that we can coordinate around or a line system that we can coordinate around and so what you run into is is lots of different standards problems again you run into companies that aren't willing to coordinate with companies and other jurisdictions you run into uh you know states that don't want to support necessarily cooperatives but will support different types of organizations like companies because they're simpler or fit better within their rules and then of course most practically you get into this problem of much larger scale

issues across States states of course being sort of the largest scale coordination we have today but many of those states don't actually have these principles in mind and they also usually fail to coordinate at a very global scale and this is sort of where oh we got actually one slide that I did not quite finish here this is uh the doing it live problem I made these last night but Austin's principles are really important there's eight of them essentially the goal of these principles is that you're creating uh local rules first and then trying to bootstrap those rules in this sort of way from this slide so that you get essentially in the end about um you know three main things happening

you have the creation of rules being very participatory within the group that the commons is a part of you have the ability for those rules to be updated According to some legible process by the community and then you have some version of uh sort of legitimacy around the enforcement of those rules which actually is very important so that people can actually get a sense of what their sort of responsibilities obligations are within those rules and what you really want is for this to be something that can be done within any kind of local organization first but then bootstrapped up in a sort of iterative way to be eventually polycentric I.E of many of these organizations many of these states companies and so forth all collaborating

together and what you get when you do that necessarily what we've seen with a lot of the Bitcoin grants rounds Bitcoin grants just being sort of a series of independent rounds that are all focused on different cause areas uh in this Republic Goods Arena so for example longevity uh things like decentralized finance things like broader artist uh Creator uh rights things like decentralized science and generally of course the three in public goods that are focused on bootstrapping open source software which is probably the one major example that's inspired me to be part of this movement in the first place we're trying to basically get all these independent rounds to start with their own rules locally start funding things

in their local communities their unshared needs using this quadratic funding idea which folks can kind of look through afterwards and the idea essentially is that we want these different uh sort of individual pieces to level up to Bubble Up from being individual human ideas that form teams that form dads that form movements to being interconnected parts of a global system that anyone can actually understand legibly and that they can believe in the legitimacy of and regardless of where they are in the world and so this is the really important I guess takeaway of this is we have a lot of things we've done really really well as a civilization most of those things have gotten is probably to about the national economy level of this little like colonic section on the on

the right here and we need to basically figure out how we're going to move from just this small sort of uh set of national economies companies which maybe even like partly multinational to a much more Network system where there is like a global set of legible rules that we can involve sort of follow and understand to be legitimate so this is again I think just some things to think about for the group as you start to build in the space if you already are building the space I hope you are already thinking about these things and I think uh that is all you need thank you [Music] [Applause]