Proposal - Request to fund FDD workstream Q4 2021 budget

đź“ś Proposals
11

I appreciate you taking the time for this detailed response. Here are my answers:

1. Remove “Community/Onboarding” and “Cross Stream DAOops” portions of this budget

  1. I believe it is an iterative process where we get better and better each round. Our outcomes have shown this (fraud tax reduction). Not only that, but we aren’t happy with just TRUSTing the results. We are taking multiple steps to VERIFY. These include adding statistical validation during GR11. During GR12 we will be starting a community run algorithm which will give more insights into the accuracy of our current process. We are also working on DAO proposals to share feature insights (for the ml pipelines) with other protocols starting with BrightID.

  2. FDD is currently managing this budget because the soft consensus earlier was that the DAOops wasn’t ready to be it’s own stream. We did not decide this, we only offered to fund and administer the funds for the meantime to make sure contributors would get rewarded.

  3. The multisig keyholders of the primary FDD multisig function more like a board as explained in our first budget request. They are not asked to partake in day to day operations. Outline of FDD Multisig Keyholder Responsibilities

  4. The process you describe that Kris brought was used to determine that FDD should continue administering this funding. This was decided by the CrossStream DAOops Governance Squad which consists of reps from all workstreams. The notes to this decision can be found here.

2. Remove “Reserve” portion of budget

We set that precedent with our first budget request and are rolling over 20k GTC of the 25K GTC reserves that we had for Q4 2021. It was beneficial for us to have the reserves this time because we were able to support the CrossStream DAOops while the situation setting it up was still somewhat chaotic. It also allowed us to fill in the gaps for the User Support functions.

Reserves for defense should be available because we don’t want to risk our ability to continue functioning based on political, technical, or other unknown issues which may arise. If this wasn’t contested before, and the outcomes have been useful to the DAO, should we not continue?

3. Focus on outcomes

FDD has a primary role that is inherently judged on outcomes. These outcomes were listed in detail both in relation to GR11 in the GR11 Governance Brief and in detail concerning Q3 Accomplishments and Q4 Deliverables in the State of FDD Q3 2021 document linked in the proposal.

In response to the specific outcome focused items you listed.

  1. Fraud tax was reduced by over 50% from $14k to $6k in GR11. But we know having one data source saying that is… questionable. Therefore we added a statistical validation which will be improved on during GR12 and another entire ml pipeline to doublecheck and compare results. It will also allow for open source feature engineering which will advance the communities ability to report mischievous behavior (Outcome #5)

  2. Collusion tax is not something we have consistently measured in the past. We agree that it should be studied in more depth. Let’s fund that research. We have two initiatives in the research part of FDD focused on this!

  3. We have consistently delivered more transparency and community involvement in the process of reviewing the grants. In addition to the ML improvements discussed above, we have also developed an appeals process that follows best practices and will be implemented for GR12.

Another piece that will help is the creation of a community data lake which is budgeted into the Data DevOps stream which we have spun out, but will work closely with FDD going forward.

Lastly, there are three open Jira tickets for the core team engineering to get to to help improve the transparency here including sending grant flags and user sanctions to a public DB similar to how new grants are handled now. We can build part of this, but we do have dependencies for the cGrants product. Additionally, it is probably best to MVP design these, test, then build a product.

  1. We began reviewing dGrants this last week. (Only 3 so far.) The curation research is closely aligned with dGrants as well.

  2. Discussed above in points 1 & 3

6.Our initiatives led by Bob Jiang in user support included creating an open source knowledge base (SourceCred) and providing translations in Chinese, French, and Spanish. This has been done in a couple months and shows the capabilities of the DAO!

In addition, the knowledge base is being spun out into the dCompass proposal at this time. User Support has a clear roadmap to it’s spinout expected in Q1 2022.

  1. This can get better. While we averaged 2-3 days during the round, post round it falls a week or two behind while we are working on the round closing priorities. There is a plan in place to improve this metric.

(This is not a fault of the FDD contributors because I personally am the bottleneck. The system isn’t built in a decentralized way yet. The contributors review, then I personally have to input this into the cGrants system. This seems reasonable at this time because it provides transparency and accountability while training new contributors during this first season)

Summary

Our outcomes are fairly clearly listed at the low level in the State of FDD Q3 2021 document. Here are the high level goals driving them.

  1. Legitimacy - Decrease the fraud tax, Establish Grant & User review process and sanctions which follow Ostrom’s 8 Principles, vigilance in assessing potential platform biases both algorithmic and human.
  2. Credible Neutrality - Increase community participation, process transparency, and accountability
  3. Sustainability - Conduct research and develop reporting to understand and communicate the threats Gitcoin faces (Grants & DAO)
3 Likes