Grants for voting neurons

Governance NNS Governance
, ,
22

Thanks for the feedback.

Furthermore, it seems we are heading towards a “bigger the Voting Power, bigger the reward” kind of mechanism.

I don’t quite understand how this would be the case or increased with having grants for multiple neurons that are decided by the community. Could you maybe elaborate a bit on why you think this is the case?
By voting on who will get these grants, couldn’t we as a community decide what we think are good known neurons independently of the voting power they already have? We could for example favour many smaller projects with less voting power or groups which we think would establish a good voting culture independent of their voting power.

I think the part about shared tools and regular exchanges is a good point! I wonder though whether this requires a certain format like one one dedicated group to arrange this for everyone or whether other formats, e.g., where multiple people contribute different things, are also conceivable. It will be interesting to see how this evolves over time!

27

These two statements seem to be in conflict. Is DFINITY asking for teams only that already consider themselves experts at reviewing a specific type of proposal, or also teams that are hoping to become experts through the process of being funded to take part in reviews?

Thanks for pointing that out! In the end the community will be voting on who gets the funds, so it is about convincing the wider community that someone is “up for the job”.

In my personal opinion, I think it is good if a candidate has the required background knowledge to at least interpret the proposals on a high level from the first week on. But they could then become more and more experts on the details of the topic and maybe be able to go deeper with time.
For example, if someone would apply for verifying upgrades, the community might want to know that this person can read Rust code and knows the IC’s basic architecture. For canister upgrades, the community might prefer to vote for people who can show that they already built or reviewed canister code before. But it might be acceptable that there is still some learning required wrt the details of the particular components or with the details of how to set up the voting neuron etc.

If in doubt, and given the information you already shard, I would encourage you to try to apply. If you are just honest about what you already know and what you still plan to learn during the process, the worst thing that can happen is that the community things the expertise is not yet enough.

What timeframe is a reviewer of Subnet management proposals expected/required to cast a vote within? (note that these proposals can appear without warning on any day of the week)

I think it would be fine to vote within the normal 4 days that the proposal is open for. Even if it is adopted early, casting a vote in the normal time would still get the neuron and its followers the rewards.
For normal (non-emergency) cases, it can maybe also be considered whether the irregular proposals can be submitted with more “warning”, e.g, always on a the same day or with a heads-up on some form thread etc. I think finding out what works best here can be one of the wins of this initiative.

Given that the level of work required to review these proposals can vary, are there any provisions for upping the funding during busy periods? (and what if a reviewer cannot keep up with a period of exceptionally high demand?)

The idea was to account for that by averaging the work hours over the weeks. It was estimated that in many typical weeks the required work would be a lot less than 2h, but as you correctly say sometimes it would go up. We tried to account for that and the overhead of just watching hte proposal action by rounding up the estimated time. We are also aware that these are just estimates

In the event that a proposal may seem to require rejection, what channel should a reviewer use to voice their concerns. Will there be a dedicated forum topic, as there is for IC-OS releases?

The idea would be to post all the verification steps on the forum. We will have to decide if it makes more sense to have a forum thread for each voting neuron or for all of them. I would then propose that concerns should be raised there.

Will there be a dedicated forum topic, as there is for IC-OS releases?

I don’t think we have decided what works best, but I like the idea! We can maybe think about what works best for the topics but also of an easy way that all the verifications of the different topics can found in a similar way!

What happens after the 6 months are up (would a reviewer reapply, or is the idea that other candidates would be given priority?)

The idea is that roughly at this point the longer-term feature, where an incentive system would be built into the system, would be implemented. So neurons that continue to be interested to be voting neurons that can be followed would be encouraged to “switch” to this system (as the design is unclear it remains to be seen what this means in detail).

I hope this answers most of your questions, otherwise please feel free to reach out again!

4 Likes
29

Thanks for the input.
We agree with this too. As mentioned in the original forum post, this is meant to be the short-term solution to already kick this off and profit from the first learnings for the different involved parties.

We will of course share design ideas for the long-term solution(s) once we have them and I propose that we then discuss the pros and cons of the different options there in more detail.

3 Likes
30

Moderator note:

@Cat5 I have removed your comments because they are frankly just antagonistic and not in good faith, nor conducive to a good dialogue or design discussion.

Please try again or refrain from such communication style.

3 Likes
31

But I’m worried about centralization. If most people delegate their votes to the big neurons to earn more rewards, it could lead to a concentration of power, which isn’t good for diversity in governance. This goes against the decentralization we’re aiming for

This is a valid concern. For the longer-term solution we can take this into account: for example, if the rewards would depend on how many neurons follow a voting neuron, we could cap that at a certain % - in that way the voting neuron would be incentivised to get some followers but also not be incentivised to get more than a certain % of followers.
I think in the current state, if there are some (up to 8) new neurons that can be followed, this would likely lead to more decentralization. Therefore, I propose that in the short-term solution we focus on engaging more neurons and then we consider these concerns for the longer-term solution.

4 Likes
32

I share the same considerations already expressed by CodeGov team members. On top of the points they’ve already raised I’m also concerned the proposed interim solution might work against what DFINITY is planning for the future, I say this based on the assumption the long term solution will be based on VP, since past discussion have hinted at it.
If in any shape or form the funding model in 6 months for such activities will be based on this principle, I believe teams are best suited to function in that kind of environment, mainly for two reasons: they are less likely to miss votes, which entices following and its more likely they can gather enough VP to sustain operations.

The problem is that the current system doesn’t account very well for teams, in fact it promotes fragmentation.
A group of 5 is incentivized to split and apply for as many grants as possible to both maximize funding and their likelihood to be chosen at all. Big teams are also completely left out of the equation.
This would result in possibly duplicating the administrative burden and surely fragmenting VP, which means once we transition to a NNS based funding model, they are less likely to have attracted enough to sustain themselves.

CodeGov is a clear example of this taking place. For the past year we have been building a public image and increasingly received the trust of stackers. With the proposed changes all the momentum will come to a halt, instead of accelerating and eventually allowing a smooth transition to the new system.

The election mechanism is also not entirely flawless, as either DFINITY will act as a filter, essentially rendering it just a grant with extra steps or the community chooses based on popularity rather than merit.

Imho a more sensible approach would have been to:

  1. Maximize interest and participation in reviewing protocol upgrades by publicly announcing grants for it, perhaps during ICP RnDs, as I think many don’t know it was something they could receive a grant for.
  2. Actively monitor the active groups and promote them to the community by explaining what they do and why it’s important, increasing the likelihood that enough VP is shifted to their neurons.
5 Likes
33

Thanks for all the reactions, applications, and feedback so far!

One of the main concerns and open questions is how this grant program can fit the larger verifying group codegov and how codegov members should apply.
I want to reassure everyone that we read all the comments and heard this feedback.

We propose to keep this forum discussion and application process open to gather more information who else applies for the grants and then discuss early next week if and how we need to reassess the details of the proposal. Several team members are also out this week and would be back by then.

Maybe codegov members could hold with applying for next week when we have this additional information? We can then also increase the overall deadline to accommodate for that if needed!

I wish everyone a great day!

8 Likes
34

Do you mind sharing your thoughts on a permanent solution? These grants are a bootstrap for sure, not a permanent solution. DFINITY shouldn’t be handing out these kinds of grants indefinitely for a variety of reasons. So how should it be funded?

@skilesare offered his opinion that it should be a penalty instead of a reward, which is a fair perspective. I would think node providers need an increase in payment if they are suddenly going to be saddled with the responsibility of securing the network or face slashing. After all, they would need to hire resources to review all the technical proposals. I guess that would be up for debate as well.

I prefer the reward approach instead of the penalty approach.

Current tokenomics and current node provider rewards do not incentivize people and organizations to review technical proposals and offer an assessment of the health of implementation of those proposals independent of DFINITY. So if we think decentralization is important, and especially if we want to reset default following with periodic confirmation of Followees and have other options besides just DFINITY to follow, then what should be the long term solution to fund this work? What is the most logical source of funding?

3 Likes
35

I would highly recommend that DFINITY creates a new post for every proposal and that the post is tagged with the proposal topic. The title of the post should include the proposal number (not just the replica version or canister ID). It is often very challenging to find forum topic that are associated with a particular proposal. The format used by the IC-OS Version Election proposal team is the most consistent and reviewer friendly in my opinion, but it could still be improved. Every reviewer could then post their review results in that post.

DFINITY should also consider giving reviewers permission to automate their posts to these threads. For example, the build / hash verification can be automated with proof of results posted automatically. When we were talking about the GaaS App (Governance as a Service) that CodeGov is working on, we wanted to produce a summary that could automatically be posted to OpenChat, Taggr, Telegram, and the forum, but we hit a road block on posting to the forum due to current restrictions. Allowing this feature could make the reviewers task much more user friendly.

3 Likes
36

If we move forward with this initiative, I would apply for Protocol Canister Management.

Also, we should probably have a forum post for each proposal for everyone to share findings and discuss changes with the team that is proposing the change.
@wpb was faster with my suggestion :slight_smile:

7 Likes
37

I don’t think there should be any grants for voting neurons—it’s like throwing money at a sock puppet!

1 Like
38

By sock puppet, do you mean you think voting neurons won’t represent real people doing real work, or that these people will be inclined to act as yes men? Either way, what makes you think that?

39

Thanks Lara, I plan to apply for Subnet management. I’ll put together an application.

I may also apply for Participant management & Node Admin. My understanding is that ‘Update unassigned nodes to latest replica’ is no longer a proposal that falls under the Node Admin topic. This came up in one of my HostOS reviews recently.

Would you be able to clarify this point in the announcement above (are certain IC OS Version Deployment proposals intended to fall under the scope of this grant or does the annoucement need amending, or have I misunderstood)?

More info here

3 Likes
40

Hi @lara,

Thank you for putting together this proposal as it is a step in the right direction. I would like to take a step back and revisit what we are trying to achieve in order to shine light on another issue.

To clearly state the goal, we would like to decentralize voting power away from Dfinity and empower individual neurons from within the community.

In order to achieve this goal,

  1. Named neurons to be incentivized to do valuable work and gain a following
  2. The IC community needs a reason to vote for named neurons other than Dfinity

Your proposal addresses point 1) above, and is a step in the right direction. If named neurons are paid, then they will be incentivized to work. That said, today, there is absolutely no reason the IC community has an incentive to vote for named neurons, EVEN IF they are doing excellent work.

I cannot overstate how big of a weakness this point is to the current design, and would like to provide an example to illustrate my point. Recently, I was debating whether to change my follow relationships for a couple of the IC topics so I could empower CodeGov. I have long admired all the work they do, especially since I know Wenzel personally, and can vouch that he is extremely reliable and devoted. I also wanted to do my small part to decentralize the IC and follow Code Gov rather than Dfinity.

All that said, I had an upside / downside equation to make. My choice was as follows:

  1. Follow Dfinity

Upside: Reliably get 100% of my rewards
Downside: The IC stays that little bit more centralized

  1. Follow CodeGov

Upside: The IC gets a little bit more decentralized
Downside: CodeGov slacks and misses a vote given its not any of their day jobs, and I miss out on my rewards

Given this upside / downside equation, I had very little incentive to follow CodeGov. As much as I want to see the IC become decentralized, it simply isn’t worth the risk on missing out on rewards.

Again, I cannot overstate how big of a problem this is to the goal of getting people to follow other neurons that are not Dfinity. We need to design a system that incentivizes folks to follow other people, where they are paid for the risk. To be honest, I am a bit surprised that this dynamic hasn’t yet been raised by the community as it seems to be very basic.

@lara I’d be curious to hear your feedback on my thoughts. While it doesn’t directly address point 1) of this topic which is the focus on this thread, I sincerely believe that point 2) is a fundamental weakness and serious time and energy needs to be spend on fixing that issue.

6 Likes
41

Thank you so much for clearly outlining both elements of this problem @dfisher . I agree point 2 is equally important to point 1. At the end of the day, there is no real reason to follow anyone other than DFINITY on technical topics even if there are known neurons that are capable and committed to offering a decentralized option. Your experience is a clear example of this point. Thank you for sharing it.

The only other example we have of trying to decentralize the NNS away from the original default following of DFINITY/ICA is the Governance proposal topic (SNS & Neuron’s Fund didn’t exist at the time and the default following when it was created followed the Governance topic, which was already decentralized). Even though the solution on that proposal topic didn’t incentivize the new known neurons via rewards, the solution to that problem did address point 2.

At the time, DFINITY was not voting reliably on the Governance topic. Hence, if you wanted to maximize rewards you had to follow either Synapse, cycledao, or ICDevs. Unfortunately, ICDevs was a little late specifically because the NNS was not decentralized and a very large genesis neuron other than DFINITY had enough voting power to block their first two attempts at becoming a known neuron…that is until Synapse and cycledao grew a following and other genesis neurons started to realize they were losing voting rewards.

The change that occurred was that the Governance topic was removed from the All Topics catch all and required explicit following on the Governance topic or you lost rewards. Since DFINITY wasn’t voting on the Governance topic, you had to choose one of the new known neurons or vote manually in order to get your rewards. The change to following patterns happened almost immediately due to the impact on voting rewards.

I’m not suggesting that this needs to be a solution to point number 2 on these technical topics. The point is just that decentralization probably requires that there is a reason to follow someone else besides DFINITY. This new grant program doesn’t address that issue like @dfisher indicated. Hopefully the long term solution that is rolled out in 6 months will include a solution to this problem. In the meantime, this grant program should offer an opportunity for known neuron to gain credibility as a Followee for these technical topics.

6 Likes
42

Yo, this is dope and has the high potential to create segmented leadership within the ICP ecosystem. Helluvah way to further decentralization. If any team or project is looking for an experienced and passionate Licensed Commercial Real Estate Broker/Professional, please do not hesitate to reach out.

44

I believe the solution lies in this direction detailed by @skilesare. It appears that we consistently face issues with improper incentivization on the IC. Despite having voting rewards, which have proven insufficient, we now have grants. What’s the next step? A larger grant?

2 Likes
45

@lara

  1. I am very interested but the compensation is quite low. The workload can vary and the compensation needs to reflect that.

  2. Also, can the engineering team provide some example reports for each task so we can understand the standard of work you require. This would really help us gauge the total workload.

  3. Also some statistics of expected proposals per month would be helpful.

  4. Does applying as an individual hurt one’s chances for selection? It is confusing on how much manpower you expect for the amount you are willing to pay.

Please provide clarity.

1 Like
46

Hey @lara and ICP community members.

I would like to be considered for the Protocol Canister Management grant, the Participant Management & Node Admin grant, and the Subnet management grant. I’ve engaged with many of you within the ICP community over the years, but for those of you whom I’ve yet to meet:

My name is Jesse D. Williams Jr.

I’m an ICP developer building Personal DAO which, in short, is a commoditized application that functions as a turn key DAO that communities can use to pool their neuron stakes, lend to earn interest, leverage their neuron stakes to borrow liquid ICP, fund community initiatives and more.

My X account can be found here: x.com
The Personal DAO X account can be found here: x.com

My LinkdIn account can be found here: Jesse Williams, M.S. - GSS Labs & Holdings | LinkedIn

My team consists of 2 developers (myself and Samuel Atwebembeire)

My relevant experience consists of:

  • 10+ years of quantitative reasoning background that includes combinatorics, optimization of algorithms for space and time complexity and teaching/training students of computer science
  • 5+ years of industry experience as a software engineer
  • 3+ years of experience developing on the Internet Computer
  • 3+ years participating in NNS governance as an active voter on the Internet Computer
  • 5+ years of rigorous independent study within the field of economics, economic history, monetary policy and capital markets

Relevant Technical Knowledge includes:

  • Familiarity with NNS architecture and functionality
  • Familiarity with the DFX SDK for building, verifying and deploying ICP canisters
  • Proficiency with Motoko, Rust, Typescript and Javascript development
  • Proficiency with Optimization of algorithmic functions with respect to space and time complexity

For those of you who would like to see a reference to my work, the best showcase would be the Personal DAO Decentralized Web Application

note: (The github repo to this project is still private pending a code audit. The link above takes you to the live instance of the first Personal DAO deployed to the internet computer. To access it, click the login button and then authenticate your internet identity. Doing so automatically submits a request for access to the DAO. Once you’ve done so, reach out to me with your principal ID and I will approve you for access to the DAO. From there, you’ll be able to participate as any other DAO member.)

My motivation for becoming a voting neuron is that I’ve invested several years of time and resources developing a digital business on top of the internet computer protocol as well as a positive working relationship with the community. I have every interest in seeing to it that the IC is properly governed so that I may continue to carry out my mission of onboarding participants of the TradFi economy to the DeFi space.

My reason for needing grant funding is that I work full time towards contributing to the internet computer ecosystem as a builder, which limits the time that I have available for pursuing other sources of income. In order to continue to contribute to this ecosystem and sustain myself, it will require some form of basic funding that allows me to meet my financial obligations in the real world.

I am acting in the long-term interest of the ICP ecosystem because this is the ecosystem that I’ve chosen to host the technical foundation of my web3 business (Personal DAO). I benefit in no way, shape or form from voting in a way that is counterproductive to the Internet Computer’s security, operations, and/or community. To do so would render my own efforts and resources over the past several years unfruitful.

I am interested in the topics for which I am applying because I understand the implications that these topics have for the development and security of the ICP ecosystem. Additionally, these are the topics that I feel confident voting on as I believe I have the technical background and experience needed in order to make an informed decision on these topics.

I will remain a voting neuron in the long term, even after the grant program has ended because, as the founder and lead developer of an ICP project, I have a vested interest in helping ensure that the network’s security and integrity is maintained for the sake of the business that I operate. Furthermore, Personal DAO requires that the DAO’s neurons select a delegate for their vote, I will be the default delegate for all Personal DAOs deployed to the Internet Computer. This means that in order to service my own community of supporters, I am required to maintain an active role as a voting neuron- independently of this grant.

My voting principles derive from a commitment to decentralization and sovereignty for users and builders- no introduction of any centralized points of failure that could be manipulated to the detriment of end-users. My conviction for this principle is demonstrated by the architecture that I’ve chosen for Personal DAO in which no entity outside of each respective Personal DAO (not even myself as the developer) may interfere with the DAO’s canister’s WASM modules nor their internal governance.

12 Likes
47

Is there some grants foreseen for just normal users trying to vote manually and trying to understand what the proposal is about?

Like some airdrop system to incentivize people to vote manually that are not tech savy… I rather see the grants/money be spend on the community and new people getting airdrops by actively participating in the voting without having a PHD in cyrptography.

From there you can then extend the airdrops for extra sidemissions on a proposal, could be even in the form of a quiz about the proposal. :slight_smile:

1 Like