Edit #2: After being informed about the current process for DAO proposals and milestones, I’m going to open a different post for dicussion based on the more current meta. Much of what is stated here is still good food for thought, but I’m going to basically launch a V2 of the discussion tomorrow.
Edit: The “act first, grant later” has apparently been used a lot, so apologies there. I’m going to make edits to cut down on what has already been accomplished within the meta, and refocus the discussion towards my proposed points of improvement
Good day to the Nervos DAO enthusiasts, today I was inspired to open a discussion for feedback on my idea for risk management regarding grant funding.
(Personal/Community introdction): For context, my name is William. I have been a prior grant recipient, but I admittedly did not succeed or properly execute the proposal. While this was unfortunate for both the community and myself, I have reflected on my failures and come to terms with it. I have consciously decided to no longer make any grant requests for the foreseable future, in order to improve my business competence elsewhere. But the actual reason I am opening this discussion is to use my failures as a lesson on how NOT to aprrove grants, and hopefully inspire better rules and regulations surrounding risk management for DAO funds.
My idea is not completely realistic due to how complicated and faulty the implementation would be in practice, but I believe the foundational thesis has merit for feedback, consideration and discussion. So without further delay, here is my idea:
This is a generalizeable framework for regulating DAO grants to mitigate the risk of negative returns on investment (the project doesn’t create a positive sum of value for the ecosystem).
The general ideas are twofold. Firstly, there could be potential improvements on how and when the funds are distributed. I got this idea from Tannr and Kevin Wang’s project Khalani, which is an intent centric project for smart contract design. This design focuses on the results of a transaction rather than the process of reaching the results. Traditionally, smart contracts are executed at the start; like the classic vending machine metaphor, you put money into the machine, and the machine then gives you the desired snack or outcome. With intents, the smart contract executes based on the results. This means that the traditional method of smart contracting gets flipped on its head, so instead of the first metaphor, you get something like this: You prove to the machine that you have the money, the machine then gives you the snack and proves that it did, now you pay the vending machine because the desired outcome is verified. @jm9k gave me a better version of this example recently: What an intent is really like is you announce publicly that you want a certain snack, and anyone in the world can compete to deliver that outcome and get paid, rather than a specific vending machine. The main point I want to emphasize is that this model is centered around the outcome. And my idea was to think about the DAO proposal and grant process in a similar way. But there are multiple models that I can derive from this abstraction:
- You make a proposal, the DAO agrees to fund it IF you are able to deliver on certain outcomes. The problem is that nobody will be able to execute the proposal without the funds.
There are multiple quasi solutions to this dilemma:
- Milestones: Let’s say we want them to commit to this proposal with a strong signal of commitment. Assume it’s a Defi project. Instead of giving the funds upfront, or dolling them out in steps, they stake their time and effort by starting the project, and doll out pieces of the funding as they hit goals. This makes the recipient sink their human capital of time, talent, and enegy into the proposal, which might help sidestep the dilemma. And it also shows a sacrifice of energy to prevent a nothing at stake problem
- Raw staking: Recipient stakes CKB or stables to get approval, and the stake is returned upon completion of milestones. The details are tricky, such as “how much do they stake” or "how to weigh the stake in relation to the value of the deliverables. This needs more consideration clearly. The other problem is a lack of capital to stake. If someone has a good idea and can deliver, but they don’t have the money to stake, this becomes another dilemma. A potential solution could be advanced staking schemes such as delegated staking, where DAO members who believe in the project will delegate their stake, and receive a small compensation from the recipient for putting up the risk collateral for them. We could also consider quadratic staking, however this method is controversial so it may introduce more complexity or risks.
- Finally we could consider an initial freebie. Traditional upfront funding, but it’s only partial. This way they can fund themselves to hit the first milestone, and then it switches to a results based model. This would basically bootsrap them to solve the chicken and egg problem with results based funding
I don’t recognize any of the solutions to be ideal or even viable. These are more like brainstorming thoughts in order to get people thinking about how to solve the bootstrapping problem with results based funding.
Another possible deisgn for the risk management framework would be even more similar to intents: anyone makes proposals normally, but only expressing a desired outcome, but anyone can deliver on the proposal, and once the quantifiable outcome is met and verified, the winner gets the reward. This is much more like a bounty board, but instead of a project listing bounties, anyone can propose one, and the DAO approves or rejects them. I doubt this idea is novel, but it’s still interesting, eespecially if we can optimize the validation, make it more autonomous, efficient, and reliable. The problem with validation is that it’s either slow because humans need to verify or dispute outcomes, or it is automated, but the automotons aren;t flexible enough to consider the nuance of non-deterministic outcomes. however, I recently discovered a project that is working on consensus for non-deterministic outcomes called “commit-reveal Pairwise Comparisons”
Anyway, these are my thoughts on how to improve efficiency, trust, decentralization, modularity, automation, and success of DAO proposals. I don’t believe that these ideas are ready for implementation. I just think as off-chain data and consensus becomes more reliable and sophisticated, while maintaining fault tolerance, and DAO governance also gains sophistication, granularity, efficiency, and reliability, these concepts may become valuable in the future.
Please provide feedback and criticism. I don’t assume that these ideas are bullet-proof. This was more a means to an end to open a discussion around risk management, inspired by blokchain technology.
-Sincerely, Will