I’ve been talking to @aiv about modifying his proposals bot to use my proposals neuron, which has 110+ ICP most of which was crowdfunded for proposals and modifying his payment mechanism so it costs only 1 ICP to make a proposal. It might be an interesting experiment to see how long the crowdfunds last and if the lower fee gets abused. Anyone can top it back up with more crowdfunds, but anyone can submit any proposal they want at only 1 ICP. This could be done without changing the proposal reject fee.
Nope I’m against it.
High entry means it must be worth it to create a proposal.
Low entry means high probabilty for spamming the nns with sinless proposal.
I think it should be 1 ICP, but I think we’d need to implement ReProposal to restrict rewards qualification to a threshold - Fix advertising Spam to enable this or we’ll just get flooded. (Or an alternate proposal that does something similar).
Unfortunately, it was voted down Internet Computer Network Status so I’d say we’re a bit away from that and clearly DFINITY has some thoughts about it, but I doubt we see any movement until after SNS:
Where we currently have concerns
- The issue currently does not seem to be urgent as we had very few spam proposals not related to financial incentives. We feel that the topic would benefit from further discussion & evaluation, compared to other alternatives (see next point). Potentially the governance working group could be a good venue for that.
- This proposal is one form of voting in stages. There are other designs in the same category, which in our opinion should be considered. For example one could consider a set-up in which
- the first phase focuses on whether a topic/proposal is worth a broader discussion
- the second phase whether one is in favor or against.
As a consequence, DFINITY opted to vote no on the current proposal. However, we would like to highlight that we believe that the proposal contains many good elements which could be leveraged for a refined solution.
I think this is not necessary right now,:
- ICP proposal submission cost equivalent to <= $100 is still considered reasonable & affordable. So I do not see any urgency to lower it now
- Anything that help to burn more ICP is desired for the health of mint/burn rate ratio.
- Can encourage more quality proposal submission to the NNS, can raise credibility & reputation of the NNS, and can reduce the time consumed by Named Neuron to due diligent for each proposals.
We should indeed lower it to 1 ICP. I like to think it’s more decentralized when its easy for anyone to participate, rather than deterring people away from it.
And yes, spam sucks, but that’s no reason to shift away from making things more decentralized. There’s currently nothing to gain from spamming the NNS, so what would even be the problem?
Edit: fixed typos.
I want to use my own neuron on the NNS to submit the proposals at 1 ICP like it was one year ago.
For every spam prevented at 10 ICP, you’re equally preventing 10 good proposals at 1 #ICP cost.
Good and bad are subjective. NNS needs a spam button. Confirmed spam proposal rewards can be restricted to 0.01 ICP - 20x less incentives for reward siphoning.
Maybe filter them in NNS UI. Set a minimum threshold - 20% or 30% to confirm spam. make sure that just two or three named neurons are not censoring proposals as spams.
Longer term, what do people think about multi-stage proposals? Something like Multi-stage Governance Proposals, Starting w/ Stage 0 and Stage 1
There could be a 1 ICP fee to submit a proposal to stage 0. Progressing to stage 1 could have a higher fee but by then the people submitting could gauge whether that seems worth it based on the feedback from stage 0.
Individual users could decide if they want to see stage 0 proposals by default, which could help filter out spam.
here is my personal feedback on your suggestion (not necessarily representing the view of DFINITY). This overlaps with points made already in this thread.
The submission cost is currently at 10 ICP in order to prevent spam, of which I see two kinds
- Skewing the reward allocation: This will be mitigated once this motion proposal is deployed. So wrt this kind we could indeed lower the submission post the deployment.
- Bad/in appropriate content which does not represent an actual proposal but uses the NNS as message board. For this kind we do not have a solution yet, but there are already some ideas around voting in stages. Thus with respect to this issue, I would not be in favour of lowering the submission cost in the moment.
There is also the additional argument that any proposal requires anyway a solid deliberation with the community. As part of the deliberation it should also be possible to find supporters to crowd fund the submission cost.
The first thought that pops into my head seeing this is - who decides what is good or bad.
ICP is 5$ right now, and I don’t think anyone would mind spending 50$ for really sending a harmful message via NNS.
This is also the reason why I suggested #5 ICP change. 5 #ICP is still 5x more than 1 #ICP cost. If we do get spammed by malicious actors, we could always increase the cost to 10 ICP if the community agrees to it.
We have decreased the governance reward back to 1. I don’t see any harm in decreasing the proposal cost back to 5 ICP and see how it goes. Then maybe to 2.5 ICP and finally back to a Neuron after we have solutions.
There is a real imperative to have more named neurons. $5 and $50 are very different most community members thinks of trying for a named neuron.
Maybe we should have Five dollar fridays or something like that where one day a week the rate is much lower…so at least if we have spam it is concentrated. We can be web3’s dive bar.
I think we should offer 20 ICP for a proposal.
- 20 ICP are given back if accepted
- If rejected, 2 ICP are burnt and 18ICP are given back.
- Third option, if 33% or 50% of voters consider it’s spam + is rejected, then the 20 ICP are burnt.
I’m sure a spam button with penalty is nearly the best way to do. Maybe I should submit a proposal
I have a question regarding the spam proposal.
Is it possible to increase the cost for a proposal if a bad actor tries to spam the nns with proposal so that the spammer pays more and more fees for trying to spam the nns with more spam proposal. The required amount after a spam can be increases in linear, in quadratic or in exponential fashion.
Maybe also reward good behaviour (who act benign) by reducing the cost for sending a proposal. To detect good behaviour an additioal button is requires (“spam”). A spam is detected if a majority has detected a spam proposal (any >50%).
So the required amount of icp at start can be set to 10 icp. It decrease for every non spam proposal by one icp until it reaches 1 icp. While in case of an spam it increases for exampe by
Exp(current icp cost + 1, 2).
Hi @Omaxus, yes that is an interesting idea. One limitation which I see is that a spamer can easily create man identities to avoid the penalty which are you suggesting.
I agree I think multi stage proposals with increasing proposal cost in proportion to stage progression would be a good fix between the high proposal costs and spamming
The penalty can be avoided through new identities, you are absolutely right and I know that.
However, if we agree that 10 icp per proposal is a good starting point for spam avoidance, then we should have some reward for good behaviour in form of proposal cost reduction (reward good behaviour) in small steps (e.g., 1 icp)
The start price for an proposal will still be 10 ICP. Which is assumed to be good enough for spam avoidance.
So if someone want to create many identies he or she still has to fill it with enough icp tokens.
- 10 identies result in 100 icp.
- 100 identies result in 1000 icp
If the one acts benigh its proposal cost decreases by one icp and in case of malicious behaviour the cost increases in a way that is not worth anymore to spam any proposal with the malicious identity, e.g., from 10 icp to (10+1)^2 = 121icp (quadratic cost function).
So If a bad actor wants to create 100 identies or more to spam the nns with 100 spam proposal or more, then it will cost him or her a lot of icp. If the icp price goes up then it become more and more unattractive to spam due to the higher cost.
So it reinforces good behaviour.
Thank you for the feedback @Omaxus !
With respect to spam due to a financial incentive:
- Post the automation of exchange rate proposals, 10 ICP is probably not good enough for a spam prevention with respect to financial incentives. See further analysis here.
- This proposal by @skilesare which was approved in the meantime entirely removes the financial incentives for spam.
With respect to spam for other reasons
- I agree that your proposal could be used in that context. And as you said, it would be more a mechanism to reward good behaviour.
- In addition we would still need to conclude what further mechanisms for spam prevention we would need (in case that 10 ICP might not be sufficient)
according to your result, I would also conclude that a solely icp token restricted spam solution is to weak.
Here is a rough scatch
- Time bound
- Reward based
- Icp Slicing
Im not sure how much effort it takes to realize it, but I think it might be a good solution because it incentives good behaviour and punish bad ones.
Every wallet can convert it rewards to proposal token. This tokens are wallet restricted. That means only the wallet which has this token can create a proposal. The tokens cannot be transfered to any other wallet.
The number of proposal token must be at first low, otherwise only people with high number of token can create a proposal. This should be avoided. It should also not to be too low at start so that a bad actors are able to create high number of wallets. It must come with a price to create multiple wallets which are able to accumulate enough tokens to spam.
If a spam proposal was detected then the wallet connected to that proposal is sliced to 50%. That means 50% of the tokens are burned.
Furthermore the cost for another proposal is increased according to an appropiate cost function (e.g., quadratic cost function). If the proposal was not a spam then its cost will be decreased (e.g., 1 ICP token).
This rough solution might reduce the number of spam per month etc. only. However, a 100% spam avoidance solution would hinder people with low icp token to come up with good ideas. We should not try to find a 100% spam solution, because it would only allow a few people to create proposal (names neuron as an example)
It also might make sense for this to scale at a constant $USD or cycles cost. Otherwise at high prices, it could be prohibitively expensive for a regular person to make these proposals, and at low prices we could see a lot of spam.
Timebound + reward based and slashing ( punish bad behaviour and incentivices good behaviour)
Are key ingredient to have a powerful spam avoidance mechanism.
Cycle cost might be a good reference value, but I wouldn’t bound it to usd. Important is, that the cost to spam should always be last attractive.
Another possible is to allow the stakers to decide whether the initial cost for a proposal shall be adjusted or not. We should not bound it to usd. Ethereum doesnt do it either regarding the required number of eth to run a node.