Deepfunding Blog

Over the past four weeks, we have engaged with community members on topics around the rules and processes of Deep Funding. In this article, we present the proposals for round 2 that have been selected for decision-making by a community vote. As outlined before the start of this process, we have filtered all submissions and ideas down to 2 questions. These are, however, not the only things we will change. We are at the early stages of Deep Funding, and it would be impractical and overwhelming to submit all possible options for change to the community. Moreover, the process of decentralized decision-making is still in its infancy, and more experience (and community engagement!) is needed to ensure that the outcomes are aligned with the majority of the community and the best interests of the Deep Funding program. 

For proper context and transparency, we will outline other rules that we are planning to add,  change, or retain as-is, and are not offered for a vote at this stage. Let’s have a look:

Some rules and conditions that we will keep

Without going into too much detail, there are a few aspects of the previous round that we are very happy with and will continue going forward. 

Grade voting

While some people used this in a rather binary way, by giving only tens or ones, we do think that this graded voting of a rating between 1 and 10 has its merits. Going forwards, we will reward voters that display a more nuanced and equalized voting behavior with extra voting weight, and this way we hope we will get even better results over time. 

Milestone-based disbursements

The one thing that we are most happy with is the method of rewarding teams on a milestone-based schedule instead of time-based. By awarding the teams for their accomplished milestones, we give them more agency in determining their optimal speed and limit the risk of awarding teams that do not deliver. Just as important, this simplifies governance, since the teams will come to us with their finished milestones, instead of having us chase them for results. This not only saves us effort but also keeps the relationships positive and constructive.

Planned changes 

Adaptations that will not require a community vote. Before zooming in on the questions that surfaced as the best or most appropriate during this experiment, let’s take a brief look at other things we are already planning to change and that, in our opinion, do not warrant a community vote at this stage. 

Separate pools for ‘Tooling’ and ‘Marketing Initiatives 

Marketing proposals are initiatives that have the goal to bring more developers or AI users to our platform. Tooling refers to functionality that aims to improve the platform itself or improve the Deep Funding processes. E.g. functionality that makes onboarding new services easier, or features that support a well-functioning reputation system. These proposal categories were already offered in round 1, but they were not presented as separate pools. By doing so we hope to bring more clarity.

• These new pools will get a bit more exposure and this may stimulate people to submit a related proposal
• It is easier to hold proposals in the other pools to the condition to develop or utilize services on the platform.
• We can predefine a better distribution of funds between different kinds of proposals, e.g avoid all rewards being eaten up by marketing proposals.

Re-distributing funds to pools that are not depleted. 

As a consequence of the above, we aim to re-distribute some funds in case they are not fully used. In other words, if we allocated 25K to marketing initiatives, but there is no marketing proposal submitted or awarded, we may distribute this amount to another pool that had more eligible proposals than available funds. The exact mechanism is not completely defined yet and is beyond the scope of this article. 

Implementing weighted votes and early reputation rating.

Depending on the outcome of this experiment we aim to start implementing the first rudimentary reputation mechanisms and also add quadratic voting weights. By adding quadratic voting weights we aim to give less weight to large wallets by using the square root of the AGIX balance as the voting power. To mitigate the undesired effects of wallet splitting, we will give additional weight to wallets that display constructive behavior on our proposal portal and proper voting behavior. The details of this are to be defined and the implementation will require some manual action in this phase. We don’t expect a perfect system overnight and in fact, do expect some attempts of participants to game the system. Nevertheless, these early measures will already create some improvements (we hope) and establish a foundation that will help us to continue to optimize the system and make it less susceptible to calculated behavior. 

Limiting the exposure of the Deep Funding treasury to extreme negative volatility. 

In round 1 we offered 1 Million USD in AGIX. The exchange rate varies, which means that the impact on our treasury varies as well. Since the disbursements are milestone-based, the impact of these awards is already spread out over time, which limits exposure to sudden spikes or downfalls. Nevertheless, as multiple rounds start to overlap we think it is prudent to define a minimum exchange rate, so our treasury will not be depleted in case of (temporary) extreme negative volatility. The exact minimum rate will be defined shortly before round 2. As with all rules and conditions, we will review the impact afterward and make changes if needed. 

Protecting the treasury this way will have some related consequences: The teams cannot be held to a promised milestone if the actual USD value of the related disbursement turns out lower than projected, due to the fixed exchange rate. To protect the teams against such unexpected budget gaps we will payout their disbursements ahead of each milestone, instead of on delivery, as we are currently doing. This will replace the current first default payout on contract signing.  Finally, we will set a guideline that the maximum milestone disbursement should be no more than 20% of the total grant, to avoid abuse, due to paying out at the start of a task instead of on delivery.
Payments ahead of milestones may also help teams ensure their contractors will be paid in a timely manner. This in turn also makes the proposition more attractive to contractors who, by virtue of the project opportunity, become introduced to SingularityNET and Deep Funding.

What we plan to do but not yet in round 2

There are some things we think would be beneficial to Deep Funding but that we are not quite ready for. The main thing worth mentioning here is ‘Expert reviews’. We think it would be helpful for the community to have proposals reviewed by experts in the field, that can assess the feasibility of a project (is it technically feasible to accomplish) the viability (What is the likelihood of accomplishing the goals, taking in to account the team and the budget) and the desirability (is there a market? What does the competition look like). These assessments would be purely advisory to the community ánd to the proposers: Reviews might lead the teams to adjust their proposal based on the advice. The end goal is not to pass judgment, but to guide the outcome of each Deep Funding Round toward proposals of the highest possible quality. 

We think expert reviews may eventually come in 2 flavors: by community members weighted by their reputation rating, and by acclaimed experts in the field, which we can perhaps recruit from previously awarded teams in the program. Both options require more completed rounds to have a sufficient number of people.

Proposals for voting

We took in all ideas from the 15 submitted proposals and filtered that down to 2 options on which we would like to consult our token-holding community. 

Question 1: Should we create a pool for searching and onboarding existing AI services?

Possible answers:

1. Yes, it makes sense to have a specific pool for existing services as an alternative to envisioned services or complete solutions.
2. Yes, but let’s not limit this to a pool, but dedicate a full round to this approach!
3. No, although onboarding existing services is important, stimulating this should not be the purpose of the Deep Funding treasury.

Context question 1:
The fastest way to seed the platform with new services is by onboarding services that already exist. We would like to motivate our community to go out and search for services that may deliver value to our platform. Likewise, we would like to motivate developers to take that extra step and onboard their service onto our platform. Some details need to be sorted out, but our conversations with the community pointed towards a similar approach as regular proposals, meaning; we will leave the proposers free in suggesting rewards and defining how these rewards will be spent. A difference is that we do see a good use case for dividing the rewards between different roles. E.g:

1. X reward for the developer for the work done
2. Y reward for the community member that discovered and reached out to the developer
3. Z reward for the community member that proactively assists in actual onboarding the service (if present)
4. Expenses for enhancing the service if needed, like making it API-ready. 

We would love to see community members go out and search for AI services, and create such a proposal themselves after they have secured the support of the developer of the service. The community member can take charge of searching and onboarding and get compensated for this. There should be some acknowledgment that the developer is cooperating. NB: All awards would only be disbursed after the service is live on the platform. 

Question 2: Should we have a small pool in round 2 just for projects in the ideation phase?

1. Yes, even though there is no immediate impact to the platform, it enables people to bring the best ideas to maturity and get early feedback. 
2. No, if the team is not able to create a proposal with sufficient detail, they are probably not ready to get funded at all. 

Context question 2:
There were multiple governance proposals that touched on the phasing of proposals and/or awarding ideas. While we do have a continuous mission on our portal for this kind of proposal, this does not always get sufficient attention. We also understand that bringing an idea to the next level may require some expenses, e.g. paying a software architect for a technical design. So this proposal would involve a new pool, just for ideas. The maximum amount awarded would be low, but proposers can get some extra attention and start forming a team this way. If we will have such a pool, we would require some tangible deliverables as output, such as a roadmap and user stories, a technical design, or a mock-up of the final solution. It is up to the teams to decide what they need to bring it to the next level, and the community to decide if it’s worth the tokens. A condition for each idea proposal would be that it aims to submit a regular proposal in the next round that will include the development or utilization of an AI service on the platform. 

Voting dates and details:

The voting event starts at the publishing date of this article and our portal will be open for a week:

• Start: October 25, 2022,
• End: November 1, 2022, 12 PM UTC
• Voting portal: voting.deepfunding.ai

We will use the method of ‘quadratic voting’ on counting the results, meaning that we will take the square root of the tokens of each wallet as the number of votes cast. With this, we aim to reduce the impact of a few wallets with very high token amounts. Considering the nature of the vote, we will not take additional measures to avoid wallet splitting. We will give extra weight to high-reputation contributors on the portal with a proper web3 login, but at this stage, the impact of that measure will be limited. The tokens should not be moved in between the time of voting and the end of the voting period. 

With this overview, we hope to give you some inspiration and motivation to support the program by giving us your opinion by means of a vote. Once the voting is done and the results are in we will prepare for the start of Deep Funding round 2!

We hope to see you all in our voting portal and on our social channels. Our community is a crucial part of Deep Funding and we value your input and ideas. Whether you engaged in the discussions on the proposals or not, we hope we can count on your support by voting for your preferred options. 

Thank you!