Increasing Decentralization on the Internet Computer

(Disclaimer: I work at the DFINITY Foundation, but did not discuss this idea with others)

The Internet Computer has received some fair criticism for its current level of decentralization due to a few large named neurons holding a disproportionate portion of the voting power.

One proposed solution to this issue is to onboard more named neurons. However any random community member becoming a named neuron is also not ideal, since governing the IC requires a certain level of technical understanding. Leading these efforts PsychedelicDAO has announced they would create their own named neuron. Besides them there is currently a vote for Crypto is Good (Daniel) whose goal with his named neuron is to encourage other teams to follow suit.

In essence we need named neurons created by competent teams of engineers who are building dapps on the IC. The problem here is that it takes a non-trivial amount of time to read, understand and evaluate proposals even if we only filter in governance proposals.

For this we need an incentive structure that achieves the following:

  • Creates a financial incentive for teams of builders to create a named neuron, participate in governance and attract a following
  • Create incentives such that these teams will vote in a way that puts the long-term success of the IC above short-term interests
  • Disincentivise popular but non-technical people from creating named neurons for financial gains
  • Incentivise investors to participate more in governance

Unfortunately I don’t know what NNS design would achieve this incentive structure, so I would like to invite discourse. Also I would like to invite the community to challenge this incentive structure or propose an alternative one that achieves decentralization and functioning governance more effectively.

That said I have a few loose pieces in mind about which I’m not very confident.

  1. To incentivise these teams to care about the long-term success of the IC they would need to lock a non-trivial amount of ICP in an 8 year neuron. Maybe in the range of 2,000 - 20,000 ICP at current prices. This is a large enough amount that would push named neurons in the direction of voting in line with the long-term success of the IC but not to dissuade teams from participation.
  2. The named neurons would receive extra maturity based on the voting power following them. This would be capped, so they can’t enter into a positive feedback loop of exponentially increasing voting power.
  3. Named neurons can follow other neurons as a fallback, so their followers would continue to receive voting rewards even in case their followed named neuron misses proposals. If a named neuron doesn’t vote manually, it won’t receive any maturity for that proposal.

I believe this is sufficiently different from Jordan Last’s proposed Committee Neurons design change, so a combination of the two should also be possible.

I very much look forward to your feedback!


Representative democracy has existed for 2500 years.

  1. Create a website where people who want to be named neuron can enter their neuron with voting history and where they can clearly state their ideals and campaign promises.

  2. Create committee and people get elected into these committee like Jordan said

  3. Allow for impeachment of named neurons so we can remove people that are causing damage.

  4. As you stated incentives wanting to be a named neuron.

  5. Keep becoming a named neuron open for anyone who becomes popular so as to not make ICP blockchain that only caters to the rich.

You don’t desincentives to become a representative you create a way to remove bad ones.

We already need to remove cycle.dao and it seems like we can’t.


So Dfinity would receive extra maturity rewards, right?

Those maturity would be taken from retails investors who are in lost 80% and are not technical enough to form a named neuron, right?

You are working for Dfinity, right?

No thank you for me.

I do work for DFINITY but did not discuss this with people from the foundation. I have added a disclaimer to the top, thank you!

I did not consider this, however I don’t think it would be a problem. At one point DFINITY stopped voting for governance proposals and removed all followers for the governance topic. The ones who wanted to continue earning voting rewards thereby started following other neurons.

I don’t have data on how many people follow the foundation on governance proposals, however because of the capping mechanism even if a lot do, it shouldn’t create too much of an imbalance. I think this would increase and not decrease decentralization.

  1. You were able to create a named neuron proposal in about a day. I think the process is no more complicated than it needs to be. If one can’t figure out how to create a named neuron, are you sure you want people to follow them?
  2. I generally like Jordan’s idea, however it’s a lot more detailed. I want to read it again and see more responses from other community members.
  3. Agreed. We need a way to remove named neurons.
1 Like

If there are going to be financial incentives for gathering a following I would consider removing the whole named neurons concept.

You get into a weird position where the current political establishment gets to decide if your new party gets to be on the ballot or not and is also financially biased towards not having you on this ballot.

And if there are financial incentives people will make sure their neuron ID is properly advertised


TBH, I know you want the best for everyone.
At this point, I am wondering why people does not use all the creative ideas to grow the network with outsiders, towards real world adoption instead of spending all that energy on not proven changes. It has been a year and half now since genesis and outside of the actual IC community, no body care about IC. The number if II account is a grossly false indication of the adoption as 90% were created in the first few months after genesis. I would be interested to know how many different II anchors are used and active the last month? I would think 5% max are active. The canisters are growing but this is inside the IC developer community. If these developers have no clients, it will be useless.

The bitcoin integration is immensely exciting for the IC community. But outside, no body talk or care about it. I am impressed for what is built on IC, like those social apps. But at the end, they don’t kill anything and don’t seems they will anytime soon. I have seen some Dfinity employees very active on Twitter but been 8 weeks and more inactive on Openchat. That tells a lot.

Why is a network like Solana, which is so far behind IC in terms of tech, has 1 year older than IC was so successful after 18 months compare to IC? Why does it have 8 times the market cap while being just 1 year older? These are the questions and where the creativity energy should be spent and stop changing the NNS and so may things that has proven nothing, make it more and more complicated and reduce the adoption.
How decentralization will be so useful if the network s not adopted?
Don’t you think the IC community and Dfinity should take some time to think about the priorities that would have a real effect on IC… adoption from outsiders, cooperation and possible of integration with other blockchains instead of fighting them, etc.
I would probable see this kind of proposal in few years, after great adoption of the IC.
Don’t think new IC clients for apps will automatically and naturally comes. We will need to work for that and put it as a priority. Right now, the priority seems to change the NNS every 2 weeks and if it does not work, we revert back. Not very productive and neither good to me.
About all the IC community says the price is not important. Fine. But the marketcap of a blockchain is a good indication of the faith of investors and adoption. 8x less then Solana, 5x less then Polkadot, 4x less then Avalanche, 2.5x less then UNUS SED LEO. (you read me right)
Hope this faith and adoption problem will be prioritized and addressed soon.

1 Like

Great point! Although it is good to have neurons named instead of people having to copy paste IDs, maybe not. Teams can have their neuron IDs on their website landing page or Twitter bio.

I understand where you are coming from and I agree with some of the points. I would like to address a few of what you said even though it’s a bit of a tangent from the original discussion.

The past year the IC community has grown immensely. Including the number of public GitHub repos, unique contributors and the IC enthusiast community. For the first year coupled with a bear market, I think this ecosystem growth is impressive. However we do keep a strong emphasis on growth and the growth team (that I’m a part of) works hard to find ways to come up with new ideas on how to grow the ecosystem.

It is not a question whether we focus on growth or decentralization. We can and should focus on both. Like it’s not a question whether we focus on BTC Integration or SNS, we have engineering resources to focus on both.

The NNS is an incredible invention that is capable of continuously updating itself. But it is imperfect and we the community should focus on making it better. The success of the NNS is largely the success of the IC, I don’t think one will exist without the other.

(Disclaimer - I am a DFINITY employee. These are my personal views, not those of the foundation)

My personal opinion is that the incentive structure already exists to encourage business owners and investors to create a named neuron and a following. Any business utilizing the IC’s tech has a risk of NNS votes negatively impacted their business (or potentially positively impacting) and thus they have an incentive to build political power within the NNS to encourage votes in advantageous ways to their self-interest. I think business owners / investors have failed to see this risk, which is why we are where we are, however the tide is moving with Psychedelic’s upcoming named neuron and conversation into the governance. All large IC companies should have a presence on the NNS, with the largest companies having their own named neuron.

Companies operating in a similar space, with similar needs/wants, could form a joint neuron to represent their common needs (similar to a lobby in politics). For example, the social dapps DSCVR, Distrikt, Taggr, Seers and OpenChat may want to form a common neuron to represent the needs and wants of the social space. Any NNS staker that believes social dapps is a high priority on the IC could then follow that common neuron.

At this point, I believe the path forward is more awareness within the business community of the risks (and possible opportunities) posed by NNS proposals and the need for them to mitigate those risks through developing a strong following of of voters within the NNS.


I can say as someone with a lot to gain if we did this, that I am 100% against this. I don’t want to make decisions on the NNS based on if my revenue stream is going to continue. We have enough of these assholes in the US House and senate.

The NNS is already designed with the right incentives. If you have ICP staked and you follow a fool(no offense to the fools court ;)) then you are going against the long term incentive of your stake.

Regarding minimums: There are lots of smart folks that are not rich and if you put a high burden on them they won’t participate. ICDevs’ treasury is currently far less than 2,000 ICP and fundraising from crypto folks experience a bear market isn’t very easy.


Haha can’t speak for the court; but none take on my part. I completely agree with you and @Kyle_Langham regarding the incentive structure.

1 Like

I agree that the minimum ICP locking threshold is a bad solution. For some, 2,000 ICP is a drop in a bucket, others will need to fundraise for it.

However I’m really not sure about the long-term success and the representation of your own team is a good enough incentive for teams to allocate time for participating as a named neuron. If it were, I believe we would see many more named neurons belonging to teams building on the IC.

But I see your point. Wouldn’t you change anything in the current incentive structure to promote the creation of new named neurons and teams allocating more time in participating in governance?

These committees built on a common goal sound like a great idea!

DFINITY stopped voting on governance proposals, but I don’t believe it is accurate to say that they removed all followers for the governance topic.

As far as I know, there is no way to remove or clear the followers of a neuron in the NNS.

Those same followers that weren’t receiving rewards after DFINITY stopped voting on the governance topic are now once again receiving rewards each time DFINITY votes on a governance proposal.

1 Like

I think compensation might be key to this. Governance proposals are probably already receiving fair compensation. It’s not too much of a burden to stay up-to-date, and the publicity and confidence of the community, while also ensuring the welfare of the IC, are real benefits.

But what about reviewing replica code changes and NNS canister code changes? This is hard work and requires a special skillset. Compensation might be key here.

Here’s a simple idea:

Manual voting gets the full staking rewards. If you follow a neuron you give them some small percentage of your rewards. Each topic can have its own “transaction fee”. Maybe governance is .25% and replica code is 1%. People are incentivized to vote manually, which increasing decentralization. But if the fees aren’t too high, then it will hopefully be worth it for people to delegate their votes to those with the resources to govern well (roadmap and code and whatever else). This will give a real monetary benefit to those seeking followers.

Can this be gamed?


I like the direction, but I’m not sure it would achieve the right outcome.

This way I might be tempted to vote manually on replica code changes, however I don’t have the time nor the expertise to review these changes. I would most likely vote with the majority (often yes). This could introduce a security risk in which someone pushes malicious code in a proposal that gets accepted by manual votes, if enough people feel the need to save that 1%. Tokenomics don’t help either, since the larger your bag, the more you care about that 1%, however the size of your bag might not correlate with your technical competency.

This would be a lucrative and sustainable business model for DFINITY to continue operating far into the future.

It would also incentivize others to become follow neurons that vote on proposals outside of governance.

However, I can definitely see many in the community pushing back against this since at least in the short term, it would result in the slowing of DFINITY’s decentralization of the IC (in terms of their ICP holdings and the resulting voting power).

1 Like

One other piece that I just realized is that yes, this can be gamed.

This financial incentive will just push developers to build NNS voting bots that will auto-manual-vote for you. A voter can then configure followees (or just vote yes/no randomly) somewhere outside of the NNS, and then run the bot at a regular cadence.


I think setting up voting committees is also a good idea, however the person or group with the named neuron will need to be able to read and understand the source code of the proposal, as well as the committee’s neurons. This board needs to have projects under construction on ICs to ensure that they are for the common good to bring about the growth of the network. because when the IC network grows, their projects benefit at the same time, creating motivation to develop ICs. these committees will resemble real-life district committees and have committee chairmen who will represent their citizens to vote. Their people can vote to remove the representative if they detect bad behavior or fail to fulfill their duties as chairman of the committee.