Feedback on the XGov Proposal

Background

  • The current initial proposal (work in progress) is here.
  • The summary of XGov session in Decipher is here.

The issues with the current proposal:

  • Likely too many XGovs, which makes the decision making process superficial, ineffective and inefficient.
  • XGovs are not necessarily qualified (in terms of having the expertise, experience, knowledge, a vision, etc.) or willing/able to commit enough time/effort (20+ hours per week).
  • No effective way for the direct collaboration (for example having weekly meetings due to the high number of XGovs) among them except probably using online forums.
  • Proposals will be limited in nature and likely won’t satisfy a unified long-term strategy. Instead a lot of time is spent discussing them, which is mostly wasted effort.
  • There is a big chance that this process causes great unhappiness and sadness in the community, and therefore leads to many people leaving Algorand for good.
  • A similar process has been used by DASH for quite a few years. The outcome has been quite disappointing (if not a disaster). For more details see this post.

Here is a concrete and relatively simple alternative proposal:

  • 5 to 10 (the exact number can be debated) XGovs are elected (on-chain voting) by the governors after reviewing their publicly shared past record and resume. XGovs are expected to be among the most qualified people in the community for the XGov role.
  • XGovs are responsible to represent governors in various decision making processes including reviewing and approving proposals/measures for governors votes. They have regular meetings among themselves, with the foundation and the community and ensure the highest priority concerns are heard and acted on.
  • XGOVs must lock $XXX ALGO each for the XGov period.

On Reddit

2 Likes

@Massimo, @Adri, @StephaneBarroso (FYI)

we should define what xGovs should do… from the definition of name it should be expert governors…

so the question is, are experts preparing docs for regular governors to vote on, or are expert governors group of people who are in “the circle of trust” to decide on more sensitive questions ?

i am worried that two level of governors will lead to inequality of people in the community and could lead to situations described in book “Orwell’s Animal Farm”

more precise argument here, is that if xGovs will decide on grants, but xGovs have just their algos locked and they do not have any income, noone will want to be xGov. Only those that want to receive funds will heavily invest into xGov so that they have better share in the decision making and can make much more money.

In my opinion algorand should create working, functional and efficiend DAO. The DAO should have under control also approval of the software commits, distribution of algos which are in the reserve, and distribution of algos which are collected from fees.

if preparation of docs for regular governor would be job of xGovs, than 5 to 10 persons is quite good. If 5 to 10 people should decide on which projects recieve grants it is not good.

my suggestion is to create delegation system where anybody can delegate his voting power to some other account in categorized manner. If I delegate my IT voting power to John for example, and he delegate his and mine voting power to Ryan, the guys with the most wisdom will vote. Note that if I vote nor Ryan nor John gets my voting power and I can do my own decisions.
This way we could calculate the 10 people with the most delegated voting power and they might be our expert governors who triggers the questions to be asked for example.

3 Likes

Yes, we need more clarification on XGov responsibilities. We can start with a smaller set like deciding on proposals and also being a more productive communication channel between the Algofam and the foundation.

The delegation system works if there are a certain set of items that you would like to vote on but if you want to discuss and make the decision on what needs to be in that list, then that doesn’t work.

1 Like

from the discussion today i think the foundation want to use it at the start mainly for distribution of the grants https://twitter.com/nonfungibleab/status/1605579457808371714?s=46&t=kPad_MguAN2dObU17YMroQ

It seems so, which doesn’t have any conflict with the alternative proposal here. I hope that instead of giving all attention and focus to the details and technical implementation of what they have in mind to the bigger picture and how the whole process effectively helps Algorand grows its adoption and community. The goal shouldn’t be to develop a multi-billion dollar game (an inefficient and ineffective process) that creates a superficial sense of decentralization but with no or very small tangible positive impact on what matters, which is the actual long-term growth and adoption.

1 Like

I think xGovs - as governance in general - will evolve over time. It’s not a trivial task to design a system that takes into account so many different opinions. The grants program is one area where results can be measured and we can use it to determine how other types of decisions can be integrated into the system.

We posted a blog on our website yesterday, announcing the creation of a governance advisory committee to give a variety of stakeholders a voice on measures and the opportunity to discuss topical issues. We are still finalizing details of how it will work and I’ll have updates in early Jan.

One thing that doesn’t sit well with me is limiting the number of xGovs. How can we make a system more inclusive by excluding people by design? I’d much prefer to have vote delegation in place, allowing xGovs to nominate peers to vote on their behalf.

We will find our way, with constructive deliberation.

3 Likes

Hi Adri,

I’d like to first thank you, Stephane and anyone else who is working on the next version of Governance.

Re the exact number of people and published doc, there is no definite magic number, it is just about how many qualified people who are willing and able to spend enough time/effort and collaborate effectively is realistic and practical. I think 14 (the governance advisory board) is not that far from 5-10 and can work in the first iteration. It seems XGovs/Advisors in the first iteration are selected by the foundation and from particular groups. To me, even I may not agree with all the details, this is not a big issue neither, as long as they are qualified and represent a broad range of opinions particularly those who may not necessarily agree with all of the Foundation decisions. The only question is how/when the information about these advisors published? And whether the process, internal discussions and meetings are becoming fully transparent? One important aspect of the effectiveness of the advisory board is reflecting the highest priority community/Algofam concerns and taking actions on them, or advising the foundation in that direction. I understand that this governance round is expected to be limited to grants initially, however they can still play this role, which is very much needed.

2 Likes

Thanks right back at you @awesoecrypto. I appreciate you and your input. Indeed a variety of voices and including opposing views is of utmost importance. We’ll disclose the advisory committees’ members in the next couple of weeks, as we prep for the first governance advisory meeting. I’m also working on adding more governance info to our website.

2 Likes

Hi Adri,

Any estimated date for disclosing the governance advisory committee members?

1 Like

Hey @awesomecrypto, I don’t know the exact date yet, because I need the committee to elect a chair before we announce. Our first meeting will happen soon and I’ll be able to share details shortly after.

1 Like