Context:
https://manifund.org/causes/manifold-community?tab=about
My plan is to submit a mechanism to prioritize feature requests and incentivize people adding those features by using bounties created by a bot. Please feel free to give any criticism or suggestions in the comments.
This system will hopefully solve two problems:
Right now, there are tons of new feature requests on the Manifold discord, which make it very hard for developers to properly prioritize them and implement the most important of those features. While I believe that Manifold users pitching in on what features would help is valuable, right now it's hard to estimate how much Manifold users would care about a certain feature existing, as the Discord upvote system is not exactly designed for that. Allowing people to vote with their mana will create a system that takes into account the strength of preferences of different people.
Manifold is open-source, but I don't think this open-sourcing has done a lot to benefit them so far. There should be opportunities and incentives for people to create their own pull requests.
Implementation (MVP)
I will be creating a bot that creates a bounty for every feature request. The bounty will be awarded to whomever successfully merges a pull request that implements the feature. When someone posts a request on Discord, the bot will copy the text into a bounty and link on Discord. The bot will add a small amount to the bounty, and the rest will be provided by users who are interested in seeing the feature happen. To claim the bounty, the feature creator would need to make a comment with a link to their pull request. I will occasionally review these award requests and grant them. Potentially this can be handed off in the future.
Potential Additions:
Message the bot to have it create a bounty for your request. The bot will also post to Discord on your behalf. Useful if you don't use Discord.
Incentivize people for bountying feature requests that get implemented, to encourage more useful suggestions, by rewarding them with mana. Potential sources of mana to give are: contributions to features that don't get implemented, a cut of the total bounty, or mana provided by Manifold.
Some ways to disincentivize controversial features, or allow "downvoting" with mana.
Open questions:
How much funding do I ask for?
What kind of collaboration with the dev team will this project require?
About Myself:
I'm a recent grad working as a data scientist. Before uni I worked in DevOps for several years. I haven't worked with the discord or Manifold APIs before but I'm sure I can get it to work.
If I decide not to go ahead and submit, this resolves N/A.
Related questions
Seems like people already did a very similar project as part of the Manifold Hackathon, and it's active now. The main difference is that they use a different website and I want to have a bot, but I don't think that's enough to justify submitting, it will be too redundant. Will resolve N/A unless somebody convinces me I should still submit this.
Manifold has mentioned that a concern of theirs about incentivizing third-party pull requests is that it will overload them with code review. To solve this, I suggest you make your bot automatically create a market for every pull request along the lines of "if this pull request is approved, will it break anything?". The Manifold devs can then safely approve any request that's at less than 5% without needing to spend much time on it themselves.
@IsaacKing Thanks for the suggestion. However, I'm not sure if such markets will have enough interest or if bettors will have enough information for this market to get a good enough estimate of risk. I will need to do more research but providing incentives and assurance for high-quality PRs is definitely a direction to look into.
@RobertCousineau I need to learn more about how Manifund works first. I still don't exactly get it. But I expect the MVP to take, very roughly, 20 hours of work, and then most of the work would be in maintaining and iterating upon this system. Because of that I would like to keep some equity in this project, but I'm not sure if it can work that way?
@Shump from my understanding, you want to keep equity in the project proportional to how much you think the benefits of the project will be seen after the fact versus before.
e.g. if it is something only you see the benefit of now, but you're sure people will understand later, go high equity. If everyone sees the benefits now, go low equity.
You could also just go with the 50/50 equity that they suggest; I'm going to do so for my first proposal.