Another possibility is for users which didn’t follow anyone to auto-reject.
Third possibility we have a prototype for is - AI voting agent. It could take the remaining non-assigned VP. In a way, if the current reviewers post their text reviews and their accept/reject votes, the AI agent can have transparent prompt and in a trustless way find out how it should vote. This will mean, there can’t be a secret agenda, so users know what they are voting for. If the prompts are set, NNS proposals have more rendered information and we have a number of voting agents held by different entities, off-chain AI can become trustless. And on-chain AI will be great at - once available. Also opens up the possibility of users having their own voting agent that uses their own prompt, not the DAO prompt in combination with all reviewers inputs.
For now, cICP will be just able to pick a neuron.
The best time to ask these questions are before they launch not after. Hopefully this topic is talked about in the node provider working group as its much harder to solve something once a product is launched.
I like the idea of auto rejecting over ai based voting.
IMO the hardest part is to prove non bias in the training of the agent vs telling others if they don’t vote it will just auto reject.
Congratulations on launching your new liquid staking protocol Infu and Borovan. I’m impressed with how quickly you have already coded and launched it. It sounds like an interesting project.
Are you aware that the node provider topology is currently full. There is no timeframe defined for when it will be open to new node providers. We currently have a large excess of nodes and adding more would mean more inflation.
Does the Anvil neuron still follow Borovan’s private neuron or has that been changed? If still following Borovan, then why not just state that you are following Borovan?
I expect this will be defined as a cluster and subject to the same rules as all other node providers. This is especially important since Borovan and the NTN dev team owns so much of the NTN voting power.
The cap of 42 nodes applies to Gen 1 node providers, not Gen 2. The Gen 2 node providers are subject to a reduction factor in their remuneration model that is sized so a node provider would actually lose money if they own more than 28 Gen 2 nodes in the best case scenario (or more than 4 nodes in the worst case scenario). There is a pretty substantial diminishing return as you approach 28 nodes that means it takes a lot longer to get a return on investment. It might not even make sense to own more than about 15 Gen 2 nodes. Of course, this is only relevant if you are able to add nodes. The NNS is not awarding remuneration to new nodes any time soon because all the decentralization targets are fully met.
Here is another one @Leadership. Why was the post above hidden automatically by the AI? It would be helpful to know what causes posts like this to be hidden. This AI moderation feature has made it nearly impossible to engage in discussion on the forum. I saw the recent post about forum rule enforcement (cross post below), but still don’t understand how that will work. The thread is locked for comments, yet it says that community feedback will be taken into consideration. I guess my feedback is that the AI sensitivity or training set needs to be updated so it doesn’t hide posts that don’t violate forum rules such as the example above.
Isn’t that counter to the basic design principles of ICP tokenomics where we want people who own voting power to have skin in the game and be subject to a minimum staking period?
In the WaterNeuron model, that voting power is owned by the WaterNeuron SNS and is triggered by owners of WTN neurons. That results in a large group of people with 2 years of WTN stake controlling the voting power of the NNS neuron that has 6 month dissolve delay.
I recommend you rethink this part of the cICP design. The voting power that is cast needs to be triggered by people who have a stake commitment that is at least as long as the neuron that is funding the liquid staking token. Otherwise people can liquid stake a lot of tokens, cast a vote, and then remove their stake immediately.
This portion sounds good until you think about it.
You are using retail funds to sell icp (constant sell pressure) While also using funds to buy nodes to increase NTN daos control over the network. NTN gets the nodes while retail gets some node rewards.
I agree that this would be a concern and have been thinking about this while eating dinner just now.
I think there’s a subtle difference in what you’ve described above and what is being proposed. I think the most important detail is the options available to these liquid stakers with regards to who they can delegate their VP to. If their only options are highly staked governance participants, then I don’t think there’s really any issue (relative to other mechanisms out there). It sounds like the details around this part of the idea are yet to be fleshed out.
This sounds like the same issue they raised with wtn. False sense of vp ownership. That resulted in claims of a “VP” grab. If they only allow you to choose from a few neurons then they control the vp the user deposits into the 8yr.
@sat you mentioned this sort of idea a while ago. Have you had more thoughts on it since then, or discussion within DFINITY about the practicalities and/or challenges?
I quoted what Infu said in response to your question. Did you see his response?
I took ‘vote’ to mean indirectly (as in by following), but maybe I misunderstood. My assumption was based on the main post - I could be wrong.
Thanks for your personal opinion. The NNS DAO is there to make adjustments of the current system and its by-laws when it’s found that the current system won’t result in a future that’s beneficial for 8 year neuron stakers - that is the primary directive that takes precedence over all others. We have found that no good will come out of the current tokenomics setup and NP topology and adjustments are needed. I can get into details if needed, but I highly doubt more explanation is needed at this point and all parties are well aware of the problems.
You can see how Anvil neuron follows Borovan (Dragginz neuron) on one of the dashboards. I think Borovan as one of the or probably the biggest 8 year neuron holder is economically incentivized to make the ICP token and NNS staking work and by doing so make it work for everyone. He spends endless hours a day reviewing proposals - every single one, not only on the NNS, but SNS as well, investigating for months and using all of his information to make the IC work. That`s aside from my personal preference in following him, after he has proven more than anyone I know that he genuinely wants the IC to succeed.
However, once cICP voting power is enough we can spend significant amount of resources on developing the new voting system and everyone can follow whoever they want.
That is still true with cICP. The value is still locked and while a few can unlock through the liquidity, not all of them can at the same time. It’s just that the locking mechanism is different.
I don’t want to get into comparing WTN and cICP, hope we don’t get there. To my knowledge, personal preference and opinion of how the NNS and voting should work - cICP design is far more superior and beneficial for everyone.
What issues did cicp solve?
From reading this post you have handed ntn dao (3 wallets) a chance to own nodes and get free vp from retail.
The tokenomics open up the retail to a bank run, they give vp to 3 people while ntn has no skin in this game.
I know a issue you focused hard on was “Skin in the game” but I can’t seem to find where you solved that, and I cant seem to find where you solved the voting centralization issues.
It’s not a personal opinion though. It’s well documented throughout the forum and takes very little research to find it. If you think you have the pull to change this situation, then perhaps you should actually raise it as at least a motion proposal. I would hate for you to have based an entire liquid staking protocol design off an assumption that you can actually add more nodes. That’s a lot of work for nothing if you can’t convince the NNS to add new inflation.
This section was a pretty long winded way to confirm that Anvil follows Borovan. Why didn’t you just state this fact in the original post?
This detail didn’t stand out to me when I read the original post. You might want to beef it up on this detail so it becomes more obvious. I suspect people who liquid stake with cICP will want to know the limits of their unstaking options.
It’s fine if you don’t want to directly compare your cICP design with WaterNeuron, but you are going to need to better explain why the NNS should be ok with cICP giving people the ability to stake and receive voting rewards that you claim will be more attractive than NNS voting rewards, yet they don’t have any time commitment whatsoever. WaterNeuron took this fundamental design principle seriously and provided a reasonable solution when they created their protocol design, but you seem to not be addressing it so far.
I don’t think I can explain it better and simpler currently, given that I am deep into the rabbit hole and some things are obvious to me as someone writing a bunch of DeFi smart contract algorithms from scratch. Please ask questions, I will answer and I think that will help in understanding the solution I am proposing.
WTN provided the polar opposite of a reasonable solution. I wish we didn’t have to fix this as well, but a fix is needed for the long term success of the IC.
@wpb I’ll be happy if we let other people speak too, then we can both resume if nobody came around to do so.
How does offering people vp with no locking required solve the issue?
Sorry, already explained
There is no unlocking. If people want to leave the liquid staking solution, they can - with their liquidity, but there are penalties if a lot of them try to do it at the same time. This de facto is a locking mechanism - just not one that is one-to-one per user, but pooled.
This ensures that most (if you do the math, possibly 95%) of the cICP votes are cast in the long-term interest of the IC.