Yep, agreed. The more layers there are, the more costly it would be for someone to take advantage of the system.
1 Like
Also, what is discussed here might eliminate one of the greatest attack vectors of the NNS. The ideal solution might make it near impossible for one person/entity to gain a majority of the voting power through the accumulation of ICP.
1 Like
My 2 cents on this:
It should be high enough that an average person (however we define that) doesn’t hit the ceiling; but low enough that the extremely wealthy individual would be extremely inconvenienced if they tried to establish multiple smaller neurons.
Yep. Would have to take that into consideration.
It seems there is consensus that having staked ICP be the main driver of voting power might not be the best governance design. On the one hand we don’t want an oligarchy. On the other hand, I don’t think folks want 1 person 1 vote either. Firstly, it is impossible in an annoymous context. Secondly, there are benefits to giving folks who have more at stake a greater say. What we want is something in between that is closer to 1 person 1 vote than the oligarchy we have today.
Yes, people parties will help. Give the average staker a lot more voting power and hence rewards power. This will help balance out the whales.
But what if we figured out a way to disaggregate voting power from staking rewards power? I am not sure why we have it as a given assumption that voting power must equal staking rewards power. Our modern society of 1 person 1 vote has nothing to do with our individual wealth. They are two separate things.
One idea I had was to reduce the voting power (but not staking rewards power) that can come from any one PrincipleID. Even though seed investors hold an enormous number of ICP my understanding is that the 48 neurons given to each investor at launch are mostly linked to 1 PrincipleID per person. So long as PrincipleIDs cannot be transferred we could if we wanted to limit folks voting power per PrincipleID. To do this fairly we would have to limit their voting power but not limit their staking rewards power.
Hopefully that was clear. And hopefully a Dfinity engineer on this thread can opine on whether what I am suggesting is even technically possible.
3 Likes
I’d also point out that any significant change like that would likely require support (a vote) from Dfinity. I doubt these whales would be willing to give up their power willingly.
4 Likes
I’m not sure what you’re referring to, a bot to do what? Proof of personhood is essential to any kind of cap on voting power, so the aim is to stop individuals from duplicating themselves through bots.
2 Likes
We absolutely need proof of personhood or we cannot implement any limits on voting power, since a user could create unlimited numbers of accounts with any amount of ICP.
1 Like
Yes that is true. My suggestion of putting a cap on the voting power (not staking rewards power) per PrincipleID without proof of personhood may still help with voting centralization for the following reasons:
-
Existing neurons. The whales today have a ton of ICP spread across many neurons but that are linked to a single PrincipleID per person. Many of that voting power is locked for 8 years and may never dissolve. Folks wouldn’t dissolve just to get extra votes in 8 years time if there was not an economic incentive to do so.
-
New Neurons. For new neurons that are created, it would be administratively burdensome for one individual to create a ton of new neurons linked to different PrincipleIDs in order to get that extra voting power. That extra hurdle may dissuade people if the economic result was the same.
1 Like
Interesting…but with scripts that administrative burden might be low enough that it would still be worth it, not sure.
I do think that keeping voting rewards equivalent to what they are today is enticing.
I’d like to offer an alternate perspective.
Decision making authority in NNS governance is not based on staked ICP alone. It is intentionally about voting power, which is derived from staked ICP, neuron dissolve delay, neuron age, and liquid democracy. Voting rewards are also intentionally weighted higher for participants with the highest dissolve delay and age, not the highest staked ICP. It is a very reasonable assumption that people with the longest dissolve delay will vote with the long term best interest of the IC because they want to make sure the decisions that are made will be in their best financial interest.
While I would prefer for ICDevs to be a named Followee at this time, I think @villa-straylight provided arguments why a group of long term investors with significant voting power believes it is not in the long term best interest of the IC at this time. They cast their votes accordingly and the NNS governance system worked exactly how it is supposed to work even though I wish the result were different.
Neuron 28 is owned by ICA and holds just over 20 total votes, yet it is by far the most influential neuron in the entire IC governance system. When ICA casts the vote for neuron 28, about 405M votes are cast automatically through liquid democracy because it has been the default Followee for All Topics since genesis. NNS decisions are not made by whales. Decisions are made by organizations that are capable of gathering sufficient voting power through liquid democracy. Or in the case of the ICDevs proposal, it was decided by a group of neurons that were sufficiently well organized and aligned on intentions to reject the proposal.
Yet only 13.5% of total voting power voted on this proposal. That is 86.5% of total voting power in the NNS that abstained. For the sake of this discussion point, I think it would be overly generous to assume that DF and ICA combined holds 33% of total voting power. If true, then that leaves at least 53.5% of total voting power that did not vote because their Followee intentionally abstained from the vote. The voting dynamics would be very different if those votes had been cast.
In my opinion, if we want to further advance NNS governance, it should come from properly incentivizing full participation in governance and from attracting new participants to NNS governance. Dfinity has already recognized this need and put forth proposals to affect this change. One of the proposals has already passed and another has been in deliberation for months. I believe the IC community needs to rally behind these ideas and recognize the governance implications they would carry.
Proposal 34485 has already passed the motion proposal phase, but the code hasn’t gone into effect yet. It has a lot of governance implications including incentivizing people to select Followees other than DF and ICA for Governance proposals. It will also enable DF and ICA to cast votes on certain Governance proposals without automatically executing liquid democracy for every NNS participant. It is the proposal that inspired cycle_dao, ICDevs, and ICPMN to make proposals to become named Followees in the NNS dApp. Once proposal 34485 goes into effect, it may only take a few additional proposals before people realize that they are not getting their full voting rewards and they will be looking for new Followees for Governance proposals. Liquid democracy voting patterns will change.
People parties is intended to attract new NNS governance participants. The most interesting part of that concept is the integration with tokenomics because it could offer very favorable incentives to a lot of new people. My perception is that the IC community so far has under estimated to potential impact that people parties could have on decentralization of the IC. We have focused too much on the mechanics of proof of humanity, whether or not a 3rd party could be developing proof of humanity, and/or complaining about the name of people parties. The reason the people parties concept is important is the tokenomics integration and the effects it could have on decentralization.
These are both really important concepts that will have big impact on the governance and tokenomics.
5 Likes
Let me make my perspective on the issue clearer. I will never vote to give huge power to an organisation basically restricted to one country. The ‘world computer’ has to mean something. Further, if that one country is the United States, that makes the situation worse from my perspective. The US has too much power as it is, since Google, Facebook, Twitter and Netflix dominate the world. With regulatory frameworks so fluid, a situation could well arise where IC Devs, beholden to US law, makes choices that aren’t in the interests of non-US ICP holders.
Poker players will know what happened on Black Friday in 2011, when the US DoJ locked down a bunch of sites, took away domain names and froze accounts, affecting every player on those sites across the globe. I can completely understand why IC Devs has proceeded in the way it has, I just want to make clear why it wasn’t only a few whales that cast votes in favour of CycleDAO but against ICDevs: it was also people like myself, though of course the big holders’ votes were the decisive ones. I do feel a sense of entitlement in some of the pro-ICDevs comments (not yours, Skilesare) on this thread, when what might be called for is more introspection.
7 Likes
Each reason listed here besides liquid democracy is derived principally from the initial ICP stake. The voting power bonus from neuron dissolve delay and neuron age are calculated from the initial ICP stake. The ICP stake is the basis for voting power, besides followers.
I do not suggest we remove liquid democracy, we want people to follow those they trust and have the ability to gain outsized voting power based on those human relationships.
I’m suggesting we cap the voting power after applying the bonuses, maybe logarithmically, and then allow the regular liquid democracy to play out.
The fundamental reason/assumption is that at some point your initial ICP stake will not motivate you to be a better voter. The current system is designed without that assumption, and I believe that will lead to a suboptimal decision-making body.
2 Likes
It is certainly anyone’s right to hold these opinions, but I’ll point out that there is absolutely no evidence that the US government has in the past or will in the future put any significant restriction on non-profits around the causes that they support…especially when it comes to technology and industry. Most of the restrictions are around limiting profits flowing to individuals in inappropriate ways. You can put as high a bar on things as you’d like, but I’ll point out that organizations like the Enterprise Ethereum Alliance are also US 501c3 organizations and supported organizations like MakerDAO and the Ethererum Foundation.
If any location inside the US is too high a hurdle, I’d encourage you to support an organization like ICDevs in a jurisdiction that you approve of. We’ll be happy to work with and pool our resources with them to move the IC forward.
2 Likes
I do believe that each person who owns and stakes ICP is generally motivated to make correct decisions regarding the IC so as to maximize the future value of their ICP. But I also believe there is a limit to the effect that a person’s ICP balance has on their ability to benefit the network through correct decision-making.
What is the difference in decision-making motivation between someone who owns 100 ICP vs 1000 ICP? 1000 ICP vs 10,000 ICP? 10,000 ICP vs 100,000 ICP?
I don’t know the answers, but I think we can agree that at some point there is a diminishing return on decision-making motivation/capability and ICP owned.
I think we should attempt to approximate this extra decision-making capability based on ICP owned.
You can read through multiple opinions on this in this thread, but basically the assumption is that wealth has diminishing effects on human beings at a certain point.
1 Like
Most people I have discussed this with have agreed with me on this.
Is there a material difference in your choices when you have 1,000,000 ICP at stake versus 1,000,000,000? I think so, but as wealth grows I do not think decision-making capability grows at equal rates. Eventually it is a diminishing return, perhaps always growing but again not by much (relatively speaking).
What is your argument against this?
2 Likes
I’m just proposing that the best way to protect ourselves from bad actors (bad decisions) is to combine the innate decision-making power of each person interested in the IC (1-person-1-vote) with the added decision-making power that their personal ICP wealth will give them (I propose some logarithmic cap).
If there is no cap, then I fear that bad actors with a lot of ICP may have outsized influence that they do not deserve, and by deserve I mean strictly in a technical sense - their voting power does not match their ability to make decisions that benefit the IC.
We want the governing body of the IC to come to correct conclusions when voting. To do this, I’m saying we should maximize the decision-making capabilities of each voter and aggregate those decisions to come to a final conclusion. Uncapped voting power per individual based on ICP stake may undermine the whole system.
3 Likes
And in case the cap I’m proposing is confusing, I’m still proposing that people with more ICP have more voting power. I’m also not proposing we get rid of liquid democracy and follow relationships, nor lockup nor age bonuses.
I’m just proposing a non-linear voting power equation based on initial ICP stake. So the more ICP you have, the higher your voting power, but also the rate at which your voting power increases would decrease as your ICP stake increases.
1 Like
I definitely agree it isn’t time yet. I just want to start getting the idea out there and gaining consensus if possible.
For the next year I imagine the community will be experimenting with various proof of personhood mechanisms. We may even see one integrated into the IC’s governance (people parties).
These developments will be very interesting and completely necessary to do any kind of vote cap.
Once we have a good proof of personhood working, I’m hoping multiple DAOs on the IC or elsewhere will experiment with various token voting models. I really think proof of personhood opens the doors to more than simple coin-based voting.
After many experiments elsewhere, perhaps it may then be time to bring what has been learned about various coin-voting methods into the IC.
6 Likes
I just want to chime in as a recent recipient of an IC Devs bounty. It seems that there is a notion out there that IC Devs has an agenda to push this experimental license on people. They’ve clearly communicated in this thread that this is simply not the case. It would be a shame to lose progress on the common goal of greater voting power distribution and more voting power for developers because of what is clearly (to me) a miscommunication.
I’m very bad about adding licenses to my work. Here’s the PR where IC devs requested to add a license to the repo they helped fund. It’s MIT.
I see that there are other topics being discussed here, but I wanted to offer a data point on this licensing issue. It seems that the ARAKME experiment is being extrapolated into “the agenda of IC Devs” by some. In my experience, no such agenda exists.
6 Likes