J2 RF 05 Nick Dodson Breaking Token Curated Registries, A Love Story

welcome everyone to breaking token curated registries so my my name is Nick I work with project called govern acts were under consensus the succubus of aetherium and we yeah we've been building a lot of governance products kind of in stealth Aragon and others know about us we talk occasionally but eventually it'll all get released and everyone be able to use it but for now I'd like to present kind of a really crazy thing that's starting to happen and and as well as show some of the weaknesses of this crazy thing and then also show people that we should be we should be critical but we should also be super intrigued by whatever the hell this thing is because it's fantastic so yep so let's get started so so before token creative registry just want to leave you guys with a you know an idea which is that human workdays will change the world so it seems a little crazy right now to say this because it's pretty early on but and the last day we tried blew up but the thing is we shouldn't we shouldn't stop there there's incredible work to be done and the governance field is really expanding and I think a lot of the projects are really traversing this space in a really interesting way so yeah I just want to leave you with that thought before we start so for the agenda what we're gonna do is I'll walk you through but it so can create a registry is I'll show you what a bit about if you were to build one right now in this room what that would kind of entail why are they useful what are the problems how can we fix those problems what can't be fixed at the moment and addressing some assumptions about these crazy down structures so so what are talking creative registries so a TCR is a decentralized way to curate a list and I'll list for any application or any protocol so it's really really really simple it's just we have this organization it's raised by a token and this token membership is going to organize itself to do one thing which is effectively curate curate a list for whatever it's supposed to be doing and hopefully do it very well so the process of doing this kind of works like this I buy a token in a crowd sale or I'm given a token all right create a token because I'm just myself and I just do this you know to myself and I take these tokens and then I make applications the applications are simple they're just I want to add X thing to this particular list and I'm going to put a certain amount of stake on this application then other token holders can come along and say sure I think your application is fine your item can go on our magical list or they can say no screw you I'm going to challenge you I'm going to put some stake against your item getting on this particular list and what I'm going to do is I'm going to say hey everyone in the token organization this application is clearly bad we should challenge it and we should remove it and so effectively we get the system where token holders now for the first time over a very poignant concept can do something with their time with their work with their labor and in this case it's curating a list and it may seem super simple but I'm gonna walk you through some examples and where this can be used and I think will totally change your mind if there's any skeptics in the room for how potentially useful this very simple idea is before that so the challenges lead to a vote if the vote passes then the application is removed or the list item is removed from the list and with many TCR designs right now we're seeing that the list item can be challenged any time so if an item gets accepted then you can challenge it at any time as a token holder so token holders are incentivized to curate the list and do it well because the value of the token is assumed to be somewhat reflective of the quality of the list in theory so individuals will want on the token because they're gonna have sway and say in the particular list they're going to be able to hold the tokens believing with the potentially valuable speculate on them etc so when it comes to building a TCR they're actually very very simple you just have an e rc 20 token and some kind of list and governance mechanic in some contract so here's just two simple arrangements of what's available for TC RS so we see in a minimal most minimal arrangement now it could actually be one contract but in a more minimal arrangement you see a token may be a registry with the governance mechanics built in in a more complex arrangement you would see the ERC 20 token registry voting mechanics and parameterization separated and with the the complex arrangement in terms of looking at parameters a lot of people like to separate their their initial parameters for the TCR to a separate separate contract so I can go into design after but I just thought I'd give because I know a lot of guys are devs like what the hell is this it's actually just a few contracts when we're getting down to the the nature of what a TCR actually is now of course you need a good concept and it needs to be worth it so why are they useful so this is I think almost more important to the whole thing we have these token organizations their doubts they're curating lists but why are these things so critical why are they so important well because we can use them for black lists white lists ICO curation there's a lot of different reasons and use cases to use these things so when we're talking about black lists there's been a big problem where meta Mask runs a with a bunch of other devs runs a black list to prevent phishing across the etherium community it's been a hugely successful project but there's one slight problem in their whole system which is their curating list and they're just deciding what they believe to be fishing and what isn't now while I love the meta math devs and there's some sitting in the room right now and I absolutely think what they believe is fishing is fishing there is a question of fairness you want a more democratic fair system to establish that blacklist you don't want it to be these specific devs that just decide with their software hey I believe this to be fishing in this isn't and so we have these cases of ethics where humans truly want to integrate some kind of democratic format to manage a particular list so we're starting to see that across the whole ecosystem everyone wants to do a TCR for something because I guarantee you if you have an application there's most likely a list and if there's most likely a list then you're most likely gonna want to do it as a TCR so you get these fantastical systems that you know didn't even have a good token model before but now they have a protocol now they have a reason to exist it is a list so so we get these magical magical scenarios now where holy [ __ ] I can do with security but it isn't security it's never security it was a protocol token it was a protocol so can all along so so big realizations are happening in the community so so then we have a few and these are ones I actually really really like so these are not people that woke up one day and said well my tokens [ __ ] so I clearly need to make it a TCR these were people that made a TCR and it was not [ __ ] so ad chain is effectively doing a TCR to decide reputable advertisers because it's a real problem in the advertising world and right now there's weird centralized systems that handle that particular kind of thing and they're ripe for disruption masari's doing a TCR that's going to curate which I cos are truly transparent truly reputable true maybe not a good investment because that would be investment advice but maybe you know well-governed in any regard and then Delphi is is arbitrating they're going to do an arbitration network and they need a list that decides hey these are the reputable arbitrators and these are not so effectively these are some really really interesting and nice use cases for curating a list that that I quite like so when we get into TC RS another thought I want to leave you with is these are safer testing grounds for doubts and for doubt technology because the TCR is not pooling a huge amount of ether and a contract and deciding where it goes we have a safe somewhat reasonable testing ground to say okay we're gonna try some doubts but they're gonna do something that's kind of soft and in the sense that it's just curating a list the worst thing that's gonna happen here is they really [ __ ] up the list which for a lot of software systems which I'll explain in a second is completely fine so that's just another thing I'll leave you with so when we talk about what are the problems with TC RS what is plaguing these organizations and I know that the the Dow memories for some of you are fresh you know it was like our 911 it was really quite quite the day so so once again these TC RS are not collusion resistant so 51% attacks and collusion attacks are going to be common and to for all you who don't know what that is you just buy the majority share of the tokens and then you just crush the list and don't think that there's people who will just buy the [ __ ] out of a crowd sale and crush the TCR because it's a good concept trolling at any level at any expense is incredibly possible so you know Barry silver it's pretty bored and definitely would be crushing some TC RS aetherium so so when we talk about coordination attacks in the current designs of TC ours no one's working through schedules so problem with human coordination that we see all the time is if you just allow things to happen at random and it will people may be asleep or on vacation or in a peculiar position where they're supposed to be doing the human work of making sure that this list is good and doesn't blow up but they're just away or not aware and so we have these really complex coordination attacks where someone may launch a very large attack you know during an event or during a particular time of year or during a particular area of time when humans are just distracted and they just don't want to touch aetherium so we get odd coordination attacks that well unlikely can work and can highly disrupt the TCR working well so even if it's not an attack to destroy it it's certainly going to disrupt it from actually performing its duty of curating the list some of you have coordination delays so basically when we talk about voting human beings are known statistically to be incredibly lazy the most lazy you can ever a dream of and and so simply some rebozo can just get through because people will not vote they will miss the boat they will simply not vote because they're they don't care or you know they don't have time or they just thought well I thought everyone else would go so just like we see today in common democracy these will plague the TC RS maybe not at first maybe when the dream is big and the list is going and you know everything's operational human beings are all excited and they're ready to vote and they're ready to try out their new democratic system and they're so happy and then just as any other democratic system before it it will slowly slowly dissipate and they will slowly not care to the point where eventually things just don't work well and the problems of democracy once again plagued another organization just this time on aetherium in tokens now that's in my opinion but I truly think it will happen so then we have continual spamming attacks so effectively just filling out the TCR with a bazillion applications now initially this may not cause too many problems but if if the TCR is designed in a particular way where applications can go in and they can be accepted without being challenged then you have a real problem on your hands because I can flood it with all these applications and people may just not sit there through time to actually challenge all the applications and actually vote on all the challenges for all the applications so you can effectively flood it completely disrupt it and maybe one of your your challenges or one of your attacks can get through now remember the applications are staked but I do believe that certain lists will be low stake list because requiring a massive stake just to get an application on a list may seem completely unreasonable that concept and those lists will be hugely targeted by this particular attack now I have fixes for a lot of these but there is one fix that is definitely not present that's definitely a difficult sell and I'll go over that in a second so so there's no clear forking mechanics as well we haven't really decided yet what is the right way to effectively copy a list copy the token copy the hold ow and maybe just not invite the attacker so when we talk about forking dad's or forking organizations we're really thinking about how can we just kind of move over here but without the [ __ ] who just rolled the DAO and so so we haven't really clearly decided yet this is the right way to do it and what's gonna happen is I think we're gonna have a plethora of potential Mechanics for this but for now some of these things are not ready to be copied they're not ready to be forked so unfortunately they will be 51% attacked or colluded against and they probably will at least the initial instantiation of the list will crumble and so the last thing is there's no clear social data to determine reasonable default parameters so we're talking about this we're talking about what happened with the DAO as well so when building dows and governance systems a big problem is setting the initial parameters what what's the base level quorum what are the initial numbers that are going to make this thing work and work well and work reasonably and so because we simply don't have a lot of social data we're going to set initial parameters that will most likely be wrong and unreasonable and it will plague the initial lists or doubts that are that are created and so when we talked about when we think about the Dow a big problem was the white paper was written it was released defaults were recommended and they were used but with no data at all to suggest why these defaults should be the way they are sure you can guess you can make a reasonable guess what is your guess good enough is it good enough to actually make this thing work now with TC RS we get an interesting opportunity to try many many many forms and many many many kinds and at a low risk for most of these people it doesn't really matter if the list works right well right now so we're gonna get a lot of incredible social data back which is actually incredibly more important I argue then most of the technology that we're building it's more interesting to see how humans are gonna work with these technologies then actually the technology itself these are pretty straightforward to build they're really simple and I bet anyone who can code here you can probably string up a pretty decent one pretty quickly so when we talk about fixing these problems there's actually pretty straightforward fixes for most of it so for 51% attacks depending on what you're doing the list for if you only use the list for subjective targets meaning the list itself is being interpreted by humans and it's being basically established as this list over here definitely is is the real one this one's definitely the attacked one so we're just going to assume this is the right list and move on with our lives when we have these subjective cases it's completely fine to have 51% attacks because effectively we can copy all the contracts we can move over here and just say hey attacker we know all your accounts because you just coordinated a massive attack against us and you exposed yourself right here so we're just gonna come over here and not invite you and it's quite simple so we can move over to a new contract but there is a problem with that for some kinds of systems but not for subjectively targeted ones so I'll get into more about that in a second so when we talk about coordination attacks if the contracts themselves implement rigid scheduling and incentivization models we're actually going to solve a lot of problems in terms of coordination right now on most of the TCR contracts you can make applications and challenges at will this is okay but the problem is if I've got all these applications coming in throughout the months then that means as a human I need to sit here all the time and monitor this thing and make sure that I'm actually doing my duty in democracy or in democratic democracies we see that we can't do this we simply can't we need to schedule hey over here is the time we do the applications and over here is the time we do the challenges that may seem simple but if you don't contractually enforce it but ends up happening is imagine if everyone in Paris at will had to go out and vote when an application came in from the government it would be ridiculous people would say are you kidding I would have to stop work I would have to leave is I you know I'm at a cafe right now I simply can't do it and so it literally crumbles so some rigid scheduling can really go a long way to help humans coordinate they know at this time they're supposed to deal with the applications and this time they're supposed to deal with the challenges now we're talking about coordination delays we're talking about things like implementing incentives to make sure that okay these people are going to vote and they're going to vote regularly and the reason why is because maybe if they don't vote they lose their tokens maybe if they don't vote they lose stake when you have things at stake people seem to do things a lot more and right now token holders what they have at stake is potentially the reflective value of their token but really for a lot of people that's a highly abstract concept they're not directly losing anything yet and I don't believe they're going to act well in an organization because of it so so definitely adding incentive mechanisms to keep participation high is one thing and on a ux side governance itself in my opinion is a lot of user experience it's basically what is the experience of running the organization what is the day-to-day experience of voting is it a game or is it literally like a Facebook you know interaction the thing is if we don't game ax Phi the user experience for governance people will not be addicted to actually participate in these systems Twitter Facebook all these complex social networks that we use all the complex apps that we use they've gamified the hell out of their app to keep participation high we can leverage this same technique in governance to make sure that human beings participate and they they want to participate so gamifying UI will go a long way in terms of making sure that we we have the the participation that we need and then lastly delegated voting I think this will be a critical marker for coordination resistance so effectively if I'm just a naive user an investor I'm not gonna want to sit here and do the voting I just won't I just want to invest in the token I like the idea but I simply will not do the voting itself this is where it's a perfect perfect case for the liquid democracy or democratic voting system so all you have to do is just have services set up when you set up your TCR that say you know what will act in a responsible way with these philosophies and these principles with these delegation principles and you can delegate to us and we'll dis manage it and the thing is it is semi centralized but it's participatory centralization it's not demanded it's opt-in so that solves a lot of horrible coordination problems when you can just delegate to a trusted party but only in the sense that I'm willing to do this so lastly we have spamming attacks which is basically going to happen regardless someone's gonna try and spam this whether it be a wealthy troll you know a very great Bitcoin er you know someone who buys a lot of ripple you know we just don't know yet who the attacker is but we could assume with these people and so so effectively we have to deal with this somehow so one way to do that is to just make the applications fail unless otherwise specified so this way they're effectively just wasting steak wasting money and you flip it on its head the reason why I say this because certainty CRS are being designed in the reverse they're being designed to accept applications unless otherwise challenged now that may seem interesting and good but the problem is is that if you want a very high-quality list and you want very high quality items you really need to know that people are going to support that they're going to add it in and there is reasons for both of these systems now in the spamming attack case if they set the stake high enough they're gonna say well they're just gonna spam it and they're gonna lose money on the staking but many lists will not be able to set the stake high enough because their use case does not warrant a 50 ether application it just doesn't it's just ridiculous so you need to when you're designing the token creator registry basically include this incorporate this thinking into the logic just to make sure that am i meeting the use case right if it's an IC o---- use case they're gonna pay 50 ether they want to be on that list they're willing to stake that whereas if it's over here and it's maybe I don't know what websites does this particular company really like that you're not gonna pay 50 ether for that it's a completely different use case so it will depend on your design so no clear clear forking mechanics if you design the list and the tokens in a copyable manner I like the minimi token where you can effectively go have contract here with state at this position we'll move over here to this new contract and we'll just reference the previous state of the old contract but we'll move on to new state new logic here then you get this nice copyable system so we can do this with tokens right now with the mini-me token which is totally ideal for governance particularly voting in contract systems but now moving forward we also need to make that list Coffeyville we need to make those items copyable so when designing these lists I think that's something to really consider and making it transitional over time and that way you can have this beautiful lineage of state and you're not too worried about you know well we've got a fork like holy [ __ ] I've got to design a new contract now just so we can move over and it's going to kill the service and kill the experience and so as for the social data experimenting heavily before production is obviously recommended but the thing is many test lists on maintenance with incentive systems in place will be really really interesting shouldn't just test on test net not with these things you want it to be as real as possible during the tests in the simulation so you should put it on main net you should test with real ether if you actually think your systems are good enough and you think it's warranted the it will not be an accurate simulation if you put it on a test that people will not act the same they will not react the same to the governance system itself so lastly what we can't fix which is address consistency cannot be maintained after a majority 51% attack so under the event of an attack the attack happens over here contract X then we move over to contract Y and that fixes the attack because effectively we moved over here and we've removed the attacker but now we have another problem we're at this address and not that address now for interpretive subjective systems or subjectively targeted systems that's completely fine we'll just interpret that this is the new system but for Kansas for systems that are contractually reliant on this particular token created registry to do a specific thing they're [ __ ] they now need to integrate a centralized body in which to interpret that's a central point of failure thus removing some of the power some of the you know the the best part about this decentralized system so this is a hard problem that exists currently with any Dow that you see now the thing is there might be solves for this eventually but right now it is extremely hard to interpret what a real Dow is and what the attack Dow is what what is the Dow that's legitimate and what is not legitimate you can try to make contractual systems you can try to make reputation based systems you can try to make identity based systems but it's incredibly difficult to decipher this particular question and it usually involves human interpretations somewhere so this particular instance can be solved I do believe it can be solved it's just I think the solve will not be exactly what we want I think it will end up becoming something of a middle ground that we're willing to accept in order to push these dowels along and use them in contractual systems so right now you know this this is something really to consider if you're thinking about doing a token computer registry and lastly when we talk about token value equaling list quality so this is a really interesting assumption and it's one that I'm just super fascinated by when Mike golden wrote the initial TCR white paper he talked about how effectively the value of the token which is this is a critical aspect of the system the the value of the token is maintained increased or reflective in some way of the quality of the list and and the thing is is is I don't know about that I mean really it would seem maybe intuitive to suggest that that is the case but in terms of economics it's really hard to derive that to be true to derive that to actually be the case what if I have a list that's incredible quality but the value continues to drop even though they continue to producing what they say does this argument still hold true now the interesting part about this is even though this is an interpretive and kind of a crazy thing can we say the same thing about aetherium is the value of that aetherium token actually reflective of what we're doing or is this the work of speculation or is it somewhere in between we really don't know but if you're depending a theorem does not necessarily depend that ether be worth a lot of money whereas the TCR is somewhat dependent within its game theoretical system to maintain to try to support that value so when we talk about breaking TCOs and we talk about what are the critical assumptions one has to make in order to make this function this seems like a very interesting attack on itself and and so I think we just want to consider this going forward I don't necessarily think this is NASA's a bad thing or whether there's a clear answer to this question but it's something you want to ask yourself both both about aetherium and both about talking creative registries as a whole so that's basically the presentation that's my Twitter these are the projects I work on the last name is my band I'm actually playing in Paris later on this is my shirt yeah okay you know it's it's great we're also playing Stockholm and London yeah so so govern X is the stealth governance project of which is not released and there's no website ejs is my aetherium JavaScript stuff which I wanted to present but I thought this is a little more relevant than JavaScript stuff consensus the the giant thing and which I've been a part of for a very long time and then check out this new module I wrote each J's extras because it's very short notation to do all the etherium stuff you really want to do it's gonna save you a lot of time if you use it but it's not in production yet so don't use it in production yet okay yeah but thank you yeah all right so how much time do we have for questions do we have time or no third time okay great okay questions yeah sorry Oh Oh when will it be I gotta get the dates it's sometime I think April something April 1st 5th area yeah I'll look it up I'll figure it out I'll post it in this slack you know yeah okay other questions okay do we have someone doing a mic or hello oh hi Frankie what is what is parallels what is parallels yeah that's the band oh yeah sorry this is really causing quite a stir isn't okay yeah I just thought I'd tag it on at the end at hobby fun yeah but okay any particularly pressing questions about you know talking governance okay okay yeah so what are your thoughts on TC R's of TC hours of TC ours right durability of alternatives it's very important for the value of tokens right yes totally so so when we when we have a TCR and in it Forks or we have a TCR that that there's maybe one toga but multiple lists exactly it is the value now completely less because we both ostracize a particular token older from the list and and now that token is still there but not worth maybe nearly as much will forking lower the value and that's maybe the target goal is to effectively just disrupt the the whole system itself to just dilute and destroy the value absolutely I think forking in this particular manner can be extremely destructive to the value and thus become a game theoretical flaw within the organizational system so I think it's time to look out for and I think it's line to account for with users with the people you're working with because if you don't you may not have a good answer for we forked six times and our token is now you know a small amount of people during this one particular list so absolutely I think it have tremendous impact on potential value there's no question yeah yeah yeah thanks for your presentation enjoyed it just a question I mean some things on internet a lot of people search for for lists yeah and I'm thinking for example the top 20 sci-fi films yeah can you see this being used in that kind of context right so so so I think I think anywhere you have a list that that human beings are in a position to say this is of great importance or this is clearly good or important I think you can definitely structure one of these organizations around now it will become so menial and so easy for someone to deploy a list for whatever you want that you could totally do a list for sci-fi films there's no this is really nothing stopping you aside from like you know 40 or 50 cents but the the trouble will be can you structure your community can you structure you know the the people around that particular topic to do this work because of a token or because they care because they want to I think there's many cases to do a TCR but never do a token sale just give it to good people and then fawcett it through those people that may totally be the right way to work these decentralize organizations sure the initial spreading of the tokens was centralized but it was to people that obviously care about this particular thing and are obviously willing to to continue caring about this thing for a long time so it will be interesting to see what we try with them but in terms of can we do a TCR for a sci-fi movie list absolutely will it work well will it be the right mechanism to do that or could they have just done that through some emails and a website I don't know I mean it sounds like maybe the case where you could just do it through some emails on a website over you know the alternative but who knows I really like I can't tell you at this time but I think it's possible there's a lot of crazy sci-fi fans out there that are willing to stake what they think is the right movie so you know like I think it's possible yeah yeah okay yep oh okay right yeah okay so the question yes yes so so I do think so to repeat the question it's just you know can we solve the address consistency problem by effectively having the owner of the contractor or a particular change in our system to to address this I think yep yes yeah yeah yeah not the not the TCR itself but but a service potentially using the TCR yeah so so in this case of course you you can structure that service contract to have an owner to have someone who's going to to determine or interpret what the the right the right now is you can do that is there is there a way once again from a data perspective from a logical computing perspective to actually determine whether before after the attack happens this is the right list and this is not it's it's difficult to say I think I think as we have identity systems basically maturing and returning and maturing over time what you might be able to say is the attacker could potentially just be a collusion Airi set of fake identities or fake accounts and you might be able to based upon a complex reputation system shift over to to the other non attack Dow and so so or sorry the the Dow which does not have the attacker so there might be some way from a data or computer science perspective to do this I think it but right now it's yet to be seen that there's an elegant system to do this or one that we can truly trust or that's easy to build without whether our existing system so that would be kind of my answer to you on that one yeah if that doesn't address it though we'll discuss it later you know I love talking to you anyway so yeah yeah I know like you know 25 percent of this room it's very funny yeah okay so other questions yeah so I have a question how do you spread new tokens properly so for example going back to the top sci-fi movies list yeah so the users that are using the tokens and voting for movies to not call in the list yeah how do you get new tokens to them and I mean if you're your user and my bad movie comes along you're sure there's gonna come a worst movie after that so there are some game dynamics going on you'll hold onto your token just so you'll be able at some later point blacklist some movie that is oh yeah totally so so there's going to be a lot of potentially very negative actors in these things as well I think there'll be a lot of people that will conduct you know methods of social manipulation and trolling later on just just to do that because because I can now I think with the voting and the challenging the applications when we talk about dispersing the tokens what we're getting into is what seems reasonable to disperse effectively the power of this organization what is the right mechanism I think for the sci-fi list not I think doing a crowd sale would be really silly I think in that case you may just want to give a lot of tokens to the top 100 sci-fi people in the world that just loves sci-fi movies the critics etc you may want to give all them tokens to start now they can give them away they can move them around they can give them to bad actors but the thing is I think in a lot of cases you're actually going to get a situation where they're gonna have a lot of fun curating that lists and arguing with each other and potentially producing this list it's what they do for their lives it is their reputation and not only that we can see all of their votes verifiably in cold blocks you know so so I think like it really really depends it depends on how you want to disperse tokens it depends on what your use cases but but in general we're gonna see some really really interesting use cases and as well potentially multi lists use cases where you know there's three or four lists one a sort reverse in order to get an application through or there's many applications for a single system and they're all working in tandem against each other potentially competing with each other to bring is a better list in the end so there's much to discover here with this particular pattern yeah if that answers your question yeah cool any other questions thoughts okay great all right thank you [Applause]