AI Assisted Minswap Catalyst Voting System - Temp Check

AI Assisted Minswap Catalyst Voting System

Table of Contents

Proposal Overview

This proposal requests DAO funding to fund completed work for the Minswap Catalyst Committee.

Our goal was to create a long-term solution that supports the Catalyst proposal selection process by improving the presentation of proposal information and simplifying the voting process. However, since it is unclear whether future Catalyst rounds will be available, this proposal only focuses on completed work that was critical to supporting the needs of the Catalyst selection committee.

We expect this work to provide additional publicity to Minswap as a leader in DAO governance with better integrations with AI.

Motivation

We made this proposal under the direction of Minswap Labs during the selection process for Catalyst F13 who said that if the work was completed without pay in the moment, they would support a proposal to fund the work.

Feedback from Minswap Labs suggested that it would be unwise for the DAO to pay for further development of the tool if it’s never used, and that additional work be compensated based on future use. As the developers of the tool, it would likewise be unwise to perform work without knowing if we would be paid for it. As a compromise, this proposal outlines payment for work completed through Fund13 to support the work of the Catalyst Selection committee.

Thus, we present this proposal for current work completed. Future work can be outlined in a new proposal when more information about Catalyst is made available or if Minswap Labs or the community decide they would like to adapt this for voting purposes.

Ownership and Open Sourcing

Minswap Labs provided feedback on the proposal regarding concerns about ownership and making sure that if the DAO pays for work completed, that the DAO should own the work. However, the DAO does not currently exist as an entity (per Minswap Labs) so transfer of ownership to the DAO cannot happen. Since it does not make sense to transfer ownership of work from one centralized group to another (i.e. we transfer ownership to Minswap Labs), we propose a second payment structure where we maintain ownership but open source the existing code under GNU GPLv3 licensing. This would permit Minswap Labs or the future DAO entity to use it. The additional payment for open sourcing is due to the opportunity costs of open sourcing the code, where Minswap Labs or someone else might continue development of the tool we had outlined in the original proposal.

Total Funds Requested

30,000 ADA - Open Sourced GNU GPLv3 Licensed

Proposal Members

Elder Millenial, Mark, and Linda are long time members of the Minswap community and key contributors to the Catalyst selection process since it began in Fund 10. They have developed AI tools, UIs, and a structured process by which Catalyst proposals are selected by a committee of Minswap community members.

Elder Millenial - Co-author of Minswap sponsored research on dynamic fees and emissions, founder of SteelSwap.

Mark, a.k.a. MeZuny - SteelSwap development partner, Cardano developer working on Eternl wallet and Iagon.

Linda, a.k.a. cryptofly777 - Fund 13 Catalyst Committee leader, DeFi Advisor, Minswap advocate.

Background

Catalyst is a system in which treasury funds are dedicated to development of new projects, and are voted on by ADA holders. Minswap, as the largest single address voter, established a DAO committee to shortlist a range of proposals. From this list, $MIN holders then vote which projects Minswap will vote for. Emphasis is put on proposals that benefit both Minswap and Cardano.

Minswap has a Catalyst selection committee that has dedicated countless hours to selecting proposals for Catalyst voting since Fund 10. Elder Millenial has built a set of AI tools and developed the original process for selection that is still being used. This has not only benefited the Minswap committee to streamline the Catalyst selection process, but also brought attention to Minswap as one of the first protocols to adopt AI into their governance process (see Links for videos of discussions).

In Fund 13, spearheaded by Mark and Linda, a new and refined method of voting was introduced after the learnings from the previous three funds. Linda refined the process, while Mark spent over 100 hours creating a web application to streamline the process and integrate Elder Millenials AI work into an easy to use interface. This dramatically improved the voting procedures and gave Minswap a MVP to become one of the largest and most transparent Catalyst voting wallets on Cardano.

Definition of Work

Mark, Elder Millenial, and Linda were heavily involved in development of the Minswap Catalyst review process for Funds 10-13. Elder Millenial created an AI ingestion and clustering workflow with new features progressing from Funds 10 to 13. For Fund 13, due to time constraints, Mark was asked to create a UI to accelerate the voting process that Linda helped to oversee.

Below are screenshots of the UI that was generated. Since it is unclear if future Catalyst rounds will exist, no additional work to any of the AI or UI systems is being proposed here. If future Catalyst rounds occur, a separate proposal may be drafted for additional features, hosting, and support of the tool.

The screenshots below are from a makeshift UI developed for Fund 13. It is not complete, and in its current form will not work for future Catalyst rounds. The screenshots are shown to provide an example of what the completed MVP will look like.

Stage 1 Voting Progress for each committee member Stage 1 Group Stage Voting on AI clustered proposals Stage 2 Individual Proposal Selection Stage 3 Group Stage Final Selections

Completion Criteria

At least 3 of the Minswap Catalyst Committee members approve release of funds based on work completed for F13.

Links

YouTube videos of public conversations including Elder Millenial discussing how Minswap has incorporated AI into the proposal selection process.

IMAGE ALT TEXT HERE

IMAGE ALT TEXT HERE

5 Likes

Hi Elder, thanks for the detailed Proposal and contributing to the Minswap DAO Catalyst filtering process. For those interested in learning more about how the process works, please see the following section in the Docs: Catalyst Voting | Minswap

As the authors mention, the Minswap Catalyst filtering process has been improving Fund to Fund as a result of learning from the past processes. At the beginning, volunteers were using a somewhat complicated system with Excel sheets, which made it harder to properly assess proposals. Recently, for Fund 13, this new streamlined process was developed. All Catalyst volunteers agree this system made the job much easier, mainly because the summaries of proposals are available directly on the frontend, and the process of voting yes or no is incredibly simple. I can corroborate that as for the first time participated in the filtering process myself as well. Wish we had a video demo where we could show everyone a bit more how the process works.

While other projects with significant impact in Catalyst voting seem to take relatively simpler approaches, with simple lists drafted by teams, or straight up aiming to financialize this aspect and sell the protocol´s voting power to the highest bidder, Minswap pulls out at all stops by designing a comprehensive community centric process. It’s due to initiatives such as these that we believe we were awarded the Best DAO Tooling & Governance award at the Cardano summit awards.

Now, getting back to the Proposal, my main issue with it is that it doesnt give any assurance of this system being used in the future in upcoming Catalyst voting rounds. Furthermore, this type of system could be expanded in its use, for other types of Governance votes. But for some reason unknown to me the authors are choosing to not open source the code but instead keep ownership of it. What exactly is the Minswap DAO paying for then? And how will this generate value for the DAO in the future?

My suggestion would be the following: the proposal authors should commit to assisting the Catalyst filtering group with this tech system in upcoming rounds. For example, a % of funds to be released after it is used in Fund 14, another for Fund 15 and Fund 16 (should be about 1 year). After Fund 16, this commitment could be renewed, pending DAO approval. This disbursement would be in addition to the 5% of the total Catalyst voting rewards authors are asking for. Linking the disbursement of funds to a valuable KPI (actual usage of the system) for the DAO that now gets a commitment for future usage in return.

1 Like

Hey @PurritoGeneral . I just edited the proposal based on our discussions.

To anyone reading this, I am writing in response to the version Purrito responded it, and also handling a couple of things discussed through Discord.

I just heavily edited the document to remove a lot of the concerns you and others in Discord have expressed. Largely surrounding whether or not there will be other Catalyst rounds. To simplify this proposal and address those concerns, we removed Milestone 2 entirely and only included payment for the current work we have done to this point. I think this is more in line with what you committed to supporting during F13.

Since we do not know if future Catalyst rounds will happen, we do not want to commit to supporting them. Especially since we don’t know what the criteria or format will be. It’s also a much larger ask to make this a more generic voting tool, since we would need to collect requirements in order to asses the amount of additional work required to complete it. Regarding the open sourcing of it, that was not a part of our original discussion, and we have priced this such that the source code would not be released. We are happy to discuss open sourcing the code or even releasing it to Minswap, but that would require us accounting for potential opportunity costs (i.e. maybe we hand it over to Minswap and they finish developing it, meaning we don’t get the work to continue developing it).

I think all of these things (open sourcing or transferring ownership, future Catalyst support, making it more generalized for DAO voting) can all be addressed in subsequent DAO proposals. After discussing with you and others, I think it is best to keep this to the work we did for F13 and you committed to supporting. Trying to add in additional things takes time and planning.

2 Likes

The edit is fair. Should be paid for work done as a sign of good faith for future prospects, but it is also laughable that this should then even be a proposal as written. It should happen merely as a matter of course and the honour of what is a handshake agreement and a verbal contract between Labs and the authors on behalf of the “DAO”. Please stop using the “DAO” as a moral fulcrum to leverage value to the protocol, which should have been generated by Labs in the first place, when it suits you.

2 Likes

As a Catalyst volunteer, I value tremendously the work this trio have put in to making the selection process feel like a walk in the park.

I’m really looking forward to seeing them work on future implementations. But before that happens, they need to be compensated for all the effort that they’ve already put in.

This team is composed of extremely brilliant minds whom Minswap cannot afford to let go. We will be very sorry if this work gets discontinued all because the DAO decided to save a few pennies (30K ADA compared to the 20M in treasury), very sorry indeed.

1 Like

Hey @theeldermillenial

I just wanted to take a moment to say how much I appreciate all the effort and hard work you’ve put into making the Catalyst filtering process easier. It’s clear you’ve put a ton of time and energy into this, and I fully believe it’s only fair for the DAO to compensate you for that.

The proposal mentions that it helped Minswap develop an MVP, which is now one of the largest and most transparent Catalyst voting wallets on Cardano. At the same time, you’ve mentioned that open-sourcing the tool wasn’t part of the original discussion.

I’m not here to question your decision to withdraw the second part of the proposal—that’s totally up to you. But it does raise a bit of a concern. If you’re being compensated for your work (which I’m 100% behind), it feels like the results of those efforts should belong to the DAO, since they’re footing the bill. And just to be clear, your talent and expertise are invaluable,I want to make sure that’s fully acknowledged.

That said, I think it would make a lot of sense to consider open-sourcing the tool, even if it’s unfinished. That way, anyone in the community who wants to use it for Catalyst or other voting-related projects can benefit from it. Plus, you could always submit more proposals in the future to refine or improve it further.

Best,
Bad-r

1 Like

I’m really sorry to see the phase 2 work taken out of the proposal. The tool that’s been built is worth twice what they are asking for already and it’s unconscionable that the DAO might do anything other than fully and fairly compensate this group. Given how much ada Minswap has requested from Catalyst, I would have expected Minswap Labs to unconditionally support the original proposal as written, and recognize it as the bargain it was.

As written, this proposal deserves full support. I suspect that Labs and the DAO will come to regret not supporting the original proposal fully.

2 Likes

“I think all of these things (open sourcing or transferring ownership, future Catalyst support, making it more generalized for DAO voting) can all be addressed in subsequent DAO proposals.”

You’re doing it again… ( we can talk about open source later… you know, in the future )

Proposal ammendments:

  1. full open source ( non starter imo ) of current work, as it is paid services ( most likely already available as I dont think there are any compiled libraries or are there? )
  2. if compensation is promised for future work, make sure the AMOUNT is handled in writing before you start work ( this protects all parties, when the waiter asks if you want mushrooms on your burger, make sure you ask HOW MUCH or you may end up with 2000$ truffles; here’s another idea; write a contract and put it on chain to the agreement, immutable… )
  3. Who will maintain this, who can understand it especially if not open source?
    –Did we create custom models using pytorch or something?
    –docs/content → send some prompts + tokens → receive summary → store results.

please understand the irony… crypto is a trustless permissionless ZKP system. yet we’re not actually using crypto to solve what is designed to solve, code is law, and having trust issues. Discussions and debates should at least in this scenario never have occurred and are obsolete.

2 Likes

Hey @Bad-r,

I just updated the document to address your concerns.

The main portion of the compensation comes from the work completed to accelerate the Catalyst selection process, which was a large undertaking on its own. We originally quoted time spent on the tool at a discounted rate, but to address your concerns we have changed the quote to reflect a fair market rate and will open source.

As mentioned in the new section titled “Ownership”, there exists no legal entity for the DAO so it isn’t possible for the DAO to own anything. As a compromise, we will maintain ownership of the code but will fully open source under GNU GPLv3.

Please do let us know any additional concerns to might have.

Best,
Elder Millenial

Hey @HaloInduction

We just rewrote to incorporate some of your suggestions.

We are now fully open sourcing under GNU GPLv3.

Under the current proposal, we do not request any funds for future work or maintenance. We are happy to review community or Minswap Labs contributions, but will not be maintaining the code in the long term unless there is an agreement in place (as suggested by your 2nd point).

No custom models were created for this. This was something we wanted to implement in the long term, but with the uncertainty of Catalyst we believe it is unwise to put that time investment into it. Right now the AI portion of this reads in Catalyst data and runs embeddings and summarizations using OpenAIs models, so use of the pipeline would require an OpenAI key. Costs for this were roughly $2 per Catalyst round.

2 Likes