DAPPCON 2019: Continuously Funded DAOs - Thomas Spofford (dOrg)

Transcript

alright hey everybody I'm Thomas I am part of Dior and we are at Dow it's focused on building developments like technical and legal infrastructure for a downs in the ecosystem I'm gonna talk about but about a potential future fundraising mechanism for douse based on the token bonding curve model I'm also gonna take my backpack off I'm not sure why I have that honkin stitch pick closer alright yeah thanks so what originally inspired us down this path of exploration is thinking about a way that dows or like decentralized organizations can fund themselves in a way that it aligns incentives across like the

different parties involved like the investors the founders and the participants of all stripes so inspired by the work of people such as simon de leur roof billy running camp and the continuous organization model the core of our approach to better fundraising is based on the toki Bonnie curve model and as we go into more detail of that I want to talk about kind of a pertinent real-life example use case of this which is the DX Dao it's live and it may want to find ways to raise money to in order to facilitate its future visions which may include both its administrative costs doing things like community engagement in marketing and development of different protocols and apps as its probably its vision or wording so the

existing solutions all kind of have limitations so I'll just mention something like a traditional equity raise even though that isn't necessarily viable in this case this has problems of both the like technical legal flexibility of distribution of equity and also as investor incentive alignment problems and that people like VCS may have a different vision for the future of the organization than the core participants in the Dow there's also an ICO type of model this works better for the flexibility distribution of equity but it has issues in terms of what utility you provide the investors beyond a speculative asset and as well as MIT maintain liquidity and market price

you have to have exchanges for that and it can be quite volatile as we've seen in the past can also forego fundraising altogether and focus on using the existing revenue which currently is for the peas from the judge ex exchange that this level of slow organic growth might not meet the needs and has a hard time like onboarding new people into the DXL community and getting them excited so meet the bonded token here this interesting icon what this is just a normal a are c20 token that is distributed by an automated market maker that gives holders rights to future cash flows and revenue from the DX downs ventures so that has those two components which all go into kind of

separately so the bonding curve is the automated market maker if you're not familiar with how this works this bonding curve contract here can mint and burn tokens in response to buys or sells and the current price is algorithmically determined by the number of existing tokens such that you know later investors pay more and it also provides a guarantee of liquidity because like all the purchases all that money that's used to purchase that stays in the curve and this is used to facilitate sells so you know that's cool but how does fundraising fit in if we're just creating this automated market maker the answer for that is by introducing a

spread to this markets so now you'll have a separate buy and sell curve whereas the sell curve is lower than the buy curve and when buys happen the income is distributed between the Dow's Treasury which is represented by the bank vaults and the reserve so it's split between these according to a parameter and this is heavily inspired by the continuous station organization continuous organization vision I recommend reading that white paper if you're interested so let's make its example a little more concrete and say this like extraterrestrial wants to invest in the DX Dow so if we would specify a collateral token when we apply this model and would say say it's died so they send and die they receive bonded

tokens in accordance to the current algorithm to determine price and that money is distributed between the Dow and the Reserve and then on cells it works much the same way separate with the lower curve you can send in bonded tokens at any time and have guaranteed liquidity according to the current price so now we have this token that's distributed continuously via an automated market maker and these sales fund the Dow while creating investors with a solidly linked incentive to the success of the Dow because they kind of are earning rights on the future cash flow as well as having the speculative potential of the asset so there's a few interesting issues here one of which is how do we handle the rights to future cash flow so

at a basic level what happens is when the Dow receives revenue that revenue is distributed according to a preset parameter between the bonded token holders in aggregate and the Dow itself so a big challenge this was probably the big technical challenge of this project is finding a way to do scalable distributions of these rewards on etherion given the current computational limitations so we have a couple schemes on this way the current one that's implemented is this Merkel tree based approach where as it works very conceptually to a stock dividend in the traditional world there's every set period a snapshot is taken of the token balances in revenue and kind of made into a Merkel tree and then this route is published by a Down member which

allows all the token holders during that period to make their their claims by using a pole model which solves a pretty cheap gas cost as well this model has some limitations obviously and that it requires the Dow to take action and it requires kind of trust that they're uploading the proper thing we generally believe that this won't be an issue in practice because it's not necessarily a deviation from the trust assumption and that they're trusting the Dow with their investment to begin with so they that kind of implies trust that they will continue to operate according to their agreements note that Dell would generally choose to maintain its reputation and the price of its asset rather than exit scam within a

single dividend period of course there are there are still the affor stated problems here so we now have believe we have a purely on chain implementation that avoids the of gas concerns with other house implementations and will avoid these concerns as well but I'm not gonna go to induct into that until it is fully implemented that's a pretty exciting development so then there's the the curve itself there's a lot of discussion along this area and so we've chosen to make a fairly generic implementation the first thing we're supporting is the bank or formula based curves which allows a pretty flexible curve parameters with very simple with a single parameter actually you can come with a pretty wide variety of schemes our ble these calculations aren't baked into the

market maker this is kind of a modular part so we allow for theoretically any type of curve to be used as kind of the lines of thinking go forward in this area the big part is also integration into Dow platforms so the DX tau is the dow stock tau and to that end we have developed a front-end within the dow stock alchemy interface and this would work just like the existing things people can make proposals in there interact with bonding curves deploy bonding curves all from a familiar face some other interesting considerations is front-running this is a problem with decentralized exchanges of all stripes

in general and we have some interesting thoughts along these lines these are different approaches that can be used we currently use limits on orders such that if the price changes significantly from what you expect it to be you can have the order fail there also is a potential to limit gas costs so that by having this fixed you can limit the potential of front-running by making it more random what's chosen first there's also a transaction batching approach such as kind of similar to the Dutch Lux exchange in the future we expect that things will be done in a more limited knowledge fashion maybe with zero knowledge proof or you could even do this with a committed reveal scheme such that order information is hidden which makes front-running difficult some future

potential features of this tool sets or things like token vesting and touching token hatching which that means that like you once you buy tokens from the curve you can't spend them immediately and this helps you know prevent pumping and dumping and create some alignment of incentives I'm on the longer term pouching just means that you only have vesting among the earlier buyers and like the vesting period kind of diminishes over time as more and more people buy there's also an idea some people like is token locking for governance rights whereas you can trade you can lock up the tokens the bonded tokens you've got for governance power in the Dow multi collateral reserves are interesting as opposed to having one currency used in the reserve this can

help with the kind of distribution of assets in their diversification might help valuation I mean we think stable coins will be prominent here anyway also potential of administration of the curve post-appointment such as the Dow could change the curve implementation if thing if kind of its assumptions are incorrect and these are all configurable is it all be like configurable things that they would choose also KYC whitelisting may be interesting things like the TPL protocol would enable this on a purely on Shan way so in summary we the continuous fundraising model we have is based on the bonding curves and you can

use this to fundraise Rudo and offer investors rights to future cash flow and with the scalable distribution if you're interested in joining the conversation you can follow this link on telegram and I wanted to give a shout out to gnosis who made this entire thing possible through their ecosystem Grants Program they've been a real pleasure to work with and they're currently accepting applications for their next wave of grants [Music] and also d org we're looking to onboard more people as we grow you know developers designers Vizio folks sales community so come check out our website at Georgia Tech if you're potentially interested in getting involved and we're happy to talk as well and yeah I can take questions yeah

there's a second mic thank you yeah just pop up your hand if you've got questions I'll start at the front cuz it's close I think it's great talk just maybe to clarify again how you handle these interest payments for the the dow token holders so you at specific intervals take a snapshot of the distribution of the tokens have stats and then how do you prove that at a later time let's say I want to withdraw my interests from the pool how do I know how can I prove that I actually received at these intervals I had this stake and that such an amount of interest was allocated to me at

different intervals so how does the proving make them is in their work sure yeah I didn't I kind of neglected to go into that well the the roots of this Merkel tree are published on chain for each interval and then the the full set of data involved is published ipfs and that's kind of assumptions that the Dow will maintain that data sets for every distribution period such as you can improve it in this case the the existence of the proof you can just submit that to the smart contract and it can verify any proof for any users allocation so once the initial route is

published anyone can withdraw on their own terms without any more human involvement they have today input like what they're wanting to claim and the proof that shows that they're allowed to claim that amount next question feel free to say wait for me if you'd like thanks again for the talk had a quick question on the on the spreads so what's the range that you first to be doing the buy and sell curves that's a good question I don't know that we necessarily foresee any range I've seen people having discussions you know all the way from the low end to the high end some people have even postulated

that there is no sell curve and using it purely as a fundraising mechanism and some people want to be something more like maybe like 10 15 percent of the my value it's totally modular within our system so we're not opinionated as to what parameters people want to choose and we feel like it's an area very like active research and development uh-huh the the question I guess behind the question was if I'm an investor right and I buy something for $100 and I can sell it immediately for $85 why would I do that right so there has to be some some really interesting expectations of ROI in the future because immediately you you're debasing the what you just put into the into the dollar right yeah you

will be having an expectation of future returns either as a speculative outside that more people will buy you'll be able to sell higher along the curve which you would get a better price than your original investment or expectation of the future cash flow improving that anyone else that had a question I have one how do you see the landscape of dolls and people playing with implementations of token bonding curves like you said it's an act like definitely things around that Clah quiddity and valuation they're active areas of research but what's going to be the difference between deals and other implementers in this space also as far as I'm aware we're the only

implementation that is focused on this kind of continuous organization inspired token bonding curve model I believe aragón black is another big one that is currently focused on more of a teiko model where you have an IC o---- and then the funds instead of being distributed to the people the members all at once are kind of distributed according to that's tap which is like a revered rate that is approved by the token holders who can vote to change that burn rate so yeah there are some different implementations going on it's it's very exciting but I believe we're the only actual implementation for this specific type if there's no one else that has questions will thank

Thomas for the presentation and I'm looking for a Jorge if anyone has seen them of the back Oh a question fantastic you course are based on bond occur on the bank worker right okay do you plan on switching on these because these are pretty inaccurate so the bank worker we just thought was an interesting one to provide some flexibility for an initial implementation we we have designed the code in such a way that the actual curve logic is modular so we're definitely interested in implementing other types of curves going forward depending on what people are interested in and what has like solid audited math on solidity and do you plan on having clothes or do you have clothes functions senior

continuous organizations do it what type of function clothes functions clothes function yeah pretty being able to tear it down being able to tear it down now we do not currently have an implemented though it would be you know a fairly straightforward thing to implement