# Taxonomy of DAOs by Kei Kreutler, Yalda Mousavinia (Devcon5)

- Channel: [Devcon](https://streameth.org/devcon)
- Date: 2020-10-09
- Duration: 38:01
- Topics: blockchain, smart contracts, distributed ledger, ethereum
- Watch: https://streameth.org/watch/yt-I2dWKKBdTAU
- YouTube: https://www.youtube.com/watch?v=I2dWKKBdTAU

## Description

Visit the https://archive.devcon.org/ to gain access to the entire library of Devcon talks with the ease of filtering, playlists, personalized suggestions, decentralized access on IPFS and more.
https://archive.devcon.org/archive/watch/5/taxonomy-of-daos

The term DAO appears here to stay. From its first use, the phrase Decentralization Autonomous Organization conjured an array of potentialities, including legal representation and management for natural systems to imaginaries of independent Skynet-like entities. However, despite the resurgence of DAO projects in 2019, we still lack a common vocabulary for different types of DAOs. 
From “human-work DAOs” to “superscalable” organizations, this breakout session aims to develop a working taxonomy of DAO projects with differing structures and objectives. We will provide a worksheet for participants to evaluate a specific DAO project instance, based on its decision-making processes, ease of entrance and exit, and broader governance mechanisms, with helpers from each project on hand to provide insight. We will also provide basis for existing organizational forms, such as co-operatives and mutual organizations, that DAOs both mimic and exceed. This breakout session seeks to establish a working group on DAO taxonomy, and chart both existing categories as well as the as yet unknown feats of DAOs.

Speaker(s): Kei Kreutler, Yalda Mousavinia
Skill level: Intermediate
Track: Society and Systems


Follow us: https://twitter.com/efdevcon, https://twitter.com/ethereum
Learn more about devcon: https://www.devcon.org/
Learn more about ethereum: https://ethereum.org/ 

Devcon is the Ethereum conference for developers, researchers, thinkers, and makers. 
Devcon 5 was held in Osaka, Japan on Oct 8 - 11, 2019.
Devcon is organized and presented by the Ethereum Foundation, with the support of our sponsors. To find out more, please visit https://ethereum.foundation/

## Transcript

[Applause] so basically the format of today's session is that you will do a brief presentation about 15 minutes or so so that we're all roughly on a similar page in terms of our working definition of those which can be iterated and changed but as well what the exercise today will look like so after that 20-minute presentation we will do an exercise to create a mappings of initial taxonomy of dials and split into smaller groups then we'll regroup and share our findings but the smaller groups again for another iteration and then we'll close with next steps for research and how to continue so the workshop is called taxonomy Dallas but it's actually towards an initial attacks on Maeve does because taxonomy is about the PhD is worth of work and we have about two hours today so we're doing a text on beat Dallas but what to tell so a lot of people use this term very differently and a lot of its origins are very fraught from where we're at today so a lot of people might remember and 2014 2015 there's a concept of kind of unstoppable businesses and also the idea of kind of runaway Skynet like entities a certain autonomous organization and the idea that it wouldn't run away there was no way either jurisdictionally or otherwise to stop it but then looking at where we are at today it does look very different than this most of them entail some form of like human interaction so for the purposes of the workshop today we put together a rough definition that's actually coming from a rough definition for audiences outside of the blockchain realm so it might not be technical as you are used to so weird money today decentralized autonomous organizations as a software based organizational framework lowering the cost of its institutional grantees and using incentive design to coordinate for common goals and of course as far as definitions go all of these separate terms inside this definition could use their own but for the purposes of today we will iterate and validating this definition through the process of taxonomy plan and get started on with the kind of definition of Dallas we thought we would highlight just a couple of dollars that I think a lot of us are familiar with that are operating today so we started to break this down into just like a few components right now to talk about but in the exercise there also we have seen more so we're prime directive that's like okay what's there's a purpose of this now and for mulligans you know accelerating the development of public aetherium infrastructure because part of the reason is yeah many teams need this throughout that they're an ecosystem but they they don't want to do it alone partly because they might think that their foundation should pay for environment but you know if you pull more hundred funds together as a network of organizations maybe we could accomplish more and then the the membership process for Maalik is an existing member of the dow needs to champion new membership proposals and as part of that championing process you have to deposit deposit ten eight and even if the member doesn't get approved in the doubt the person that champions it will get the ten each so there's like slashing if you propose a bumper that's not that that didn't get admitted and then basically the voting process for all proposals whether it's for becoming a voting member or applying to get shares that you prefer them so you can develop something you just need like that a simple majority and there's no foreign required so if Moloch is about sixty people now like you know five people show up and in the vote it will it would still pass and as I said before it's like there are two types of proposals but as far as the smart contract goes it doesn't really care and here are applying to be a member or plan to be a grantee of the project and then Moloch also has an exit process not all now's necessarily do but it's a novel ragequitting mechanism where where it allows you to exit the doubt and especially if a proposal is about to be passed and you don't align with it or you feel that that dowel is moving in the wrong direction it's like wasting money you don't want it to take some of your money you can reach but and the proposals are in a grace period you'll get your proportion of the Treasury back so so yeah Moloch is unique in the sense that there is a common Treasury and each person owns a stake of it because whenever you're applying for membership here you're putting your depositing money into the doubt in exchange for your shares so that's basically describing my life briefly so how many people are two million with Genesis tell Chrissy's joke ends okay so about a third of people so Genesis now is another example of a doubt that's operating today it's based on the Dow stock framework and operated by Dow stats how many of you familiar with dowsed active decision hands so if you are so you can kind of and then might disagree with some of what I say that this is what I think please man I see Genesis as the main kind of advocacy tool for Dow staff so it started by Dow staff members and it's often maintained by larger kind of Genesis deaf community so similarly in keeping and defining it now having a prime directive Birds kind of objective function to simplify for the Genesis tau which is built on a dowel stock framework its role is to accelerate the development of public Dow stock based infrastructure and organizations so that means the Genesis now is primarily to support both the technical development of Dow staff but also the community community and social development of now stock so what it looks like in practice is like a community forum that you use men in masks or a nurse's station to interact with in some transactions and you come to the decisions about who might be funded for some community work community fundraising who might be funded for some technical work like building bonding curbs for down but it looks mostly like before I've got people vote on and come to conclusions about how to support doused out another larger frame so the membership process basically it works stupid for different - cocktails but for the genesis doe it's more culturally determined at the moment so the kind of cultural norms around becoming a member of the Genesis Dow is that a wallet address ideally a human identifiable person but that's not technically hard coded submits a proposal advocating for their own membership so what this looks like I did it and there is basically a Google form that I put together it being like my name is Kia I want to join the dog because I want to support this community doing XY said here's what I've done in the past to support this community please accept my request for reputation so reputation is not to be confused with Al Gore's reputation it's actually a token that's internal to dowse cocktails and is not transferable so you can think about it almost in a sense of um you might have a token that's financial capital whereas reputation and it does satisfy social capital so when you advocate for your membership what that means is that you are given reputation or social capital by the DAO and thus have voting power in the down so they're members of process just to simplify as you go to the Dom say hey I want to join and if people vote on whether you join or not and usually you're using some sort of persona or human-readable identity okay for yourself as the voting process is somewhat complex but to do a super simplified version this is based on the Dow stack framework to genesis now once again is based on a Dowsett framework and they use holographic consensus but you can basically say there's like two tiers of how the governance process works one is that there is a basic voting process but on top of this voting process there's independent prediction markets where you use the Gentile set up in to predict whether a proposal will pass or not if you predict successfully that a proposal will pass it will get boosted in the main interface and being boosted this means that kind of appears at the top of the interface and no longer requires an absolute majority to pass so absolute majority means on 50% of the people who have voting power need to vote instead it has relative majority which means that only 50% of the people who vote of the reputation voting power that votes has to vote in order for it to pass so basically to simplify that there's a prediction market that says I think this proposal will pass independent of the boat and then if that predictor market does it means it's a lot easier for that book to either pass or fail if you don't have a written proposal it's really hard to get something passed um simplified in there's a lot of other kind of smaller nuances but I don't want to go necessarily into them in detail but very happy when we go into breakout sessions to help any group that wants to look at this stuff and one thing that's interesting in terms of preparing Genesis to optimal lockdown is the exit process so where is Moloch I think a lot of us are familiar with the kind of grapes quitting everyone likes using the term in endow styles you could say that the exit process is actually technically undefined whereas in Mohawk it's leverage and in Genesis you can say it's actually difficult to exit the doubt you could do a proposal that says hey I want to be you no longer recognized member but since at least kind of technically and speaking their reputation or voting power is not transferable there is no way that you should be able to give it up unless you were to in some way and maybe like publish the private key or something so there's not really a way that the exit is technically leveraged or socially leveraged and a sec - yeah um so for today um to switch gears a little bit those are some down really super brief examples but bring it back to text on amines aside and to look at where we are at with what we mean about to us today we're building on the work of this paper and hopefully you guys can paint it um it's called a method for taxonomy development and its applications and information systems its primary authors Nickerson and it was published in 2012 um so we wanted to use a little bit more rigor in terms of how we approach this rather than being like these just fit in this group these other down-spin another group so we're gonna do is another super brief run-through of the methodology will be following today before we break out into groups and try it ourselves so to define taxonomy now taxonomy is a system that describes how different concepts are related and organized within a specific hierarchical structure but in the context of this workshop and the contents of the paper that will be following they define rather laborious ly a taxonomy as a set of dimensions consisting of mutually exclusive and collectively exhaustive characteristics so what does that mean mutually exclusive means that any goes once we have two worksheets on everything it will become more clear but mutually exclusive means that one doubt in a certain dimension can't belong to multiple different characteristics so you can't say that at AU if you look at membership entrance process you can't say that it does both permissioned and unless you have to pick one of the other and the taxonomy categories have to be mutually exclusive in that way and now can't belong to several of them at the same time otherwise you won't produce a kind of clear tree of a taxonomy does that make sense okay and collectively exhaustive characteristics every job has to there shouldn't be a characteristic that's empty or there's a big job that doesn't fit on the list because then the taxonomy doesn't include the whole field and you could say that it's used to explain and operationalize the category of objects a taxonomy is so this means that it shouldn't necessarily be perceived as an ontology it could be used to produce an ontology but mainly and throughout the research paper it says what's a what's a useful touch on me with a good taxonomy and the tautological definition is that it's one that is used so we don't think this is necessarily a finite kind of description of a static field but one that's put into use to produce new language to produce to also highlight where there are gaps currently so this is a diagram of our flowchart Larina go through was in that paper and it basically describes the process so first to begin with you'll determine the Medicare characteristics so it's a high level interaction between for our case it's a high level interaction between a user and a decentralized autonomous organization and then the ultimate like what it what are our ending conditions and it's to conclude and an initial taxonomy you will map a set of downs and then mutually exclusive and collectively exhaustive characteristics so that's essentially with Kea just went over so this is like different different paths you can take you can either start with conceptual to empirical basically you can conceptualize new characteristics and dimensions of doubts examine the objects for these characteristics and dimensions and then create or revise your taxonomy and I fir empirical two conceptual it's identify a subset of cows I died by common characteristics and group the dowse and then group these characteristics in two dimensions to create or revise your taxonomy so as you're going through this process like I mean we did exercises before we're talking about jowls we're figuring out the characteristics and then that's was flowing into applying them to other cows and it's just like this loop until you like come up with like the the best terms to describe or to create the taxonomy so so as so as we said in the definition of taxonomy it's a set of dimensions used to used to consist used consistently mutually exclusive and collectively exhaustive characteristics so what is the dimension what is the characteristic so you have one dimension you can have multiple characteristics so one example is like she just said membership entrance permissioned permissionless closed we can talk more about you know what clothes might be I don't know if we want to talk about that now but that's kind of like it's like for permissioned it's there's like a clear way like an encoded in the smart contracts for how you can enter or permissionless it's like that's also in the smart contracts but you don't have to like apply to join you can like by the token on the market and then for clothes that's this kind of like it's a decentralized autonomous organization but it's it could be something like with ten member so you're like a ten people you're in a group of you just kind of like decide one to add people to the dowel or not just like based on the people that are in the organization so that's more of a Dow and the technical sense of like people are creating an organization on the blockchain but it doesn't really have like easy and true paths or yeah essentially so then basically what we're going to do in developing the taxonomy is we come up with more dimensions and we come up with more characteristics like whenever we went through that example with Moloch and Genesis before we also defined something like exit process etc etc so after this after this present presentation is done we're going to do some of these exercises ourselves but basically in the worksheets what we're going to be doing do you want to talk about this here yeah so that was our kind of brief presentation and to give an overview of taoism taxonomy design so do we feel like but okay so let's say what we're gonna do first and then we'll see how clear it is so what we would like to do is actually have people break into groups of it's quite a lot of us so it may be five to six people um with the people around you and we are going to pass out worksheets where we've actually predefined dimensions and characteristics for the taxonomy such as medicines membership entrance membership exit the number of active participants legal entity etc um and so it's a kind of gritted row like this and all of these that are on the slide right now so and we've suggested several example doubts for you to evaluate in your group so we suggested Moloch Genesis vivo McMurdo and aetherium and of course some of these may or may not be what we come to in the future but you are also welcome because there are blank rows and columns in the worksheet to choose your own does to evaluate and also to suggest your own dimensions and characteristics that should belong then taxonomy based on the brief framework we just laid out so for the moment we will take about 15 to 20 minutes in groups - and we ideally have each group evaluate 2 to 3 Dow instances and it is encouraged for you to pick tabs that are not on the worksheet and actually on you know I'm happy to provide and go around and do a list of dowels that we're more familiar with and I'm happy to help you work with them and what I think will be just roaming to provide any missing information may bring up the spreadsheet so another thing that we can do together before we start going through before we start doing this for in the tabs on here the the dimensions that we've defined we can also come up with a view as a group that can be written on the sheets just so we can go through the process of you know defining the dimensions and characteristics together so maybe we can just start off by doing one and see like right now like right now if someone wants to raise their hand in 4 they're dimension but we have years membership entrance membership exit number of active participants legal entity dispute revolution live resolution my life single-phase human activity and deliberation it's like there are a few more dimensions and if you have an idea and raise your hand right now so when you say guess legal guess like it has a legal rapper type like for example like me for doubt there are a lot of stakeholders like bash back foundation or Aragon Network Tao I didn't put on that sheet because I don't know how many people are with the operations of Aragon dal but I love John to put it on the sheet so so yeah that's that's kind of what we wanted to means but for now we can see if anyone has an idea for another dimension and if you don't we can get engagement or desire and so we have number of active participants and the way that we have defined that right now is how many addresses or nodes have participated in governance over the past twelve months so we tried to define that as like engagement but I guess that those threshold would be like how many keep how many people are members versus how many people are active that might be a phrase it better yeah yeah I'm thinking about aims of having a high participation rate over to submission port city okay so it's not really as to what's happening but this design alright let's let's go throughout notice that this will be interesting to see if we can create like some dimensions around it so maybe let's try treat I like five or ten minutes just talking here on this use case I'm in so what would you what would you define as the characteristics of this hold okay I see where you're going with this gasps say that again yes yes hey okay that's the closet okay and then what's the next so now we're going to no wonder what are the other characteristics are enragement something how many times a day should you check how much daily attention to inquire or like looking at sir seen with any about SARS or that compost so no we're defining yeah I need your help here I'm not I'm not seeing these us as related yeah are you oh I was just struggling to separate two types of thousands that once one percent participation in one that was thirty percent so I'm kind of like warm or let's say attention hungry doubts and attention and white house that's what I was trying to separate yeah yeah so how would you and you're gonna take Moloch and Genesis as an example version no pity so he where it's a when it's on the sheet right now so the two options are assumption hungry well ideal ideal governance through low percentage subset of users IPO governess through high percentage I guess the problem I find with is it's like it's like a very it's not a significant amount of data to be able to set what that threshold parameter be so you would want to look at all existing and this is why I realize it's kind of like a PhD approach everyone is that you would want this data to be able to say like the threshold of activities like 30 percent or like how many transactions interact with today's smart contracts on a given day and you would want to look at all of the operating dose and then find an average there and choose the break point so it's really hard to quantify for this started say like there's a quorum which is like level of engagement and kind of mandated technical engagement in order for the doubts of function but then there's also just amount of transactions interacting with the smart contracts of the job on a given day and you could but also I think that there is a little bit of a chicken and egg problem where you could say that oh it's optimal governance is for a small amount of people but then it's like what if that optimal or stated objectives change over time and then it becomes for a large or whatever they're both kind of in a recursive let's say that you make in your next on unit now stays their desired then actually going to deploy actually structure shapes agency agency shape structure I think it would be really cool from we try to figure this one out and in does that specifically they're thinking off creating like upgrade your down buttons and missus options would be something easy to deploy on that you would have like ideal governess for fifteen people later down reaches 25 people there's a bug like hey do you want to take a moment to reflect on different parameters yeah that's very interesting then I and we can talk more about maybe an example where it might be easier that's not the worksheet that could be kind of related just like generally what is the voting process you know is it is it based on one person one vote you know like you have some kind of identity mechanism so you can actually use that governance process as a token wait is it like reputation weighted so it's like that that's an example of something that we could have as you know how do people engage with the voting process because if you're talking about because things upgrade that is introducing another like voting mechanism right so then that mechanism can just start to be defined so I mean I think that could be an interesting exercise or dimension to think about in this worksheet because that is like also one of the most like easier ones to grasp but yeah I think the ones that involve data are super Innis interesting but difficult to do in the context of this workshop like we're even thinking as far as like okay how do we like some of these maybe it's like a Gaussian distribution and we want to like figure out what's that number they're like what's that number therefore like average holdings of the part of the average like current average number of tokens at each like a theorem address it's like well maybe you have like you know a lot of people at the top or maybe it's not like that it's like a different distribution curve so I think there are a lot of like interesting metrics that would be nice and apply on them see where the difference the decision is that made on general change you know to be these paths the episode is to repeat that so that everybody heard what refers to mention it does the dominant ones or not yes or no and then there would be a corollary to that was if the management of funds is directly on chain so a doubt could be used just to make decisions about where fights go but it doesn't necessarily facilitate the actual transaction which is something that is low degrees in funds but the villains inventing a policy in a company right what organization make sense so it could also just be implementing a policy we can definitely selling um so that it is certain things of like it does get a little ambiguous here because with the theory of most interaction to involve some sort of funds in some way always but we could say is there a smart contract that holds manages a comment funding pool and it meant is it that funny pool man is watching through the doubt decision-making process great that everybody was potentially use the common you know dimensions that are you looking for or classifying because that's any morning so is it this position well all right does it have like like these requirements right is it is it is it because a required full circle for me personally as an ancient house right are their guarantees right a serious and what some of attack vectors you know - so people definitely use those I think what we set out to find is the meta characteristic for gymnasts autonomy is on a higher level human interactions in the dough and a lot of those definitions are very much on that kind of protocol like layer 0 level bond so I think that they I would definitely encourage we can use some of them liveness asynchronous yes no he's happy to write to use those as write a bit but I think that we are not necessarily doing a kind of base layer examination along those lines sneaking like this monetary revocation hybrid pie I recognized one of the ones are thinking and son is it profit or nonprofit but I like the distinction I've done also a good but like for profit hierarchical for profit cooperative for content you could make a smaller distinction at a more granular distinction there tonight and then for security token we can do like like for the ones that we don't like people created them and they never really said what it was and it's funny like exchanges to analyze those but there are like you know we're total best say this is a see very totally maybe no it's going to be harder on also have you like you somebody yeah but people might say but it's curious so you can say utilities okay they say it's one but maybe not like that that means like you know another branch it's like is it no because the team settings nowhere because it was analyzed by someone else now this might be too low-level but maybe like which like framework is going on totally and you could say also like little tiny framework or not don't worry and realize this is all sort of like kind of Turtles all the way down because you whatever yeah well I mean you could say like everything based on Aragon OS it's either compatible with Aragon west or it's not sounds like and probably the same with like this are the lowest level statutes and you can also say like Forks that follow up all right I'm not saying they're not know yet you can say that they are built on existing yeah [Music] yeah I caught it by that small can I see what you mean now every like on what depends a lot on what the search is also to repeat so it was right up jean-yves consensus before an option happens or can it an action be initiated before consensus is reached so you could use a concrete example of can the can some of the funds in the Commons pool be distributed without having however consensus is defined in the Dow how before that's reached so you could say kind of consensus or core independent action like maybe also later that would it make more sense to have like that list true or false it's associated like does it use whatever like Robin like if you use that for which system does it use and then you end up having like a really long list of options and it might be crossovers when it comes to taxonomy what's the matter - have it your falls to the long list or to have a single category for options so I think I think this is one of the like and as you do this like true or false it might be like part of that recursive that kind of like loop or your like pain maybe there's a better way to group this singer with all these yes or no so I think that that's kind of like the process like it might be easier to start with like stinking yes/no hybrid but it might just like this might just fit more into like okay well what's the voting process because you know it can go really deep like mammals you know like yeah that's exactly what's what's going on here so we can like start to name different may not like the ultimate goal is see if we can give each doll like one like name it's like Moloch is but like just come up with we're going to end up committing all the same crimes of two nails you know in the end I feel like my my each that's starting with white hairs for a symbol it permissionless next on me that requires I mean I think part of the problem is like you say decentralized autonomous organization people don't really know what that means so it's like it's really bad yeah it's like I don't think that's gonna make people realize like what these organizations are is just like I'm a corporation it's like okay well you know a see your your cooperative really good so we can all benefit provision are you benefit yeah so it's like that's what's gonna make evolve and the doubt ecosystem but since into the characteristics or more likely hold and corporate long that's going to be like a little bit different or by be more types of ways that were classifying things and yeah just like in your corporate bylaws you might have like different characteristics of how you might allocate funds or what your purposes but it still kind like rolls up into a simple simpler definition yeah so this was our exercising dimensions and so you're to see on the worksheet said there's a lot of blank spaces so feel free to add some of these in classify or if there was something that you really thought that we didn't discuss that you want to use to classify our taxonomy is to print out she'll figure out that yes yeah we didn't add that but you there are blank spaces so you can add we didn't really define everything we wanted to leave some of the earth for us all to decide and then it's like how many groups so we're gonna have about 30 minutes for this and then we're going to put them into the spreadsheet and map and know what language is all producer