Tokenomics Proposal [Community Consideration]

Thanks for this @MichaelAnnh. It sounds like the key point of difference right now comes down to the actual risks and concerns posed by enabling neuron transfer - we both agree that neuron transfer would be great to enable, so long as it’s safe (read: secure against 51% attacks, and generally still conducive to long-term good governance of the IC).

Are you on Telegram? If so, might I suggest that we connect there to discuss - don’t want to air possible attack vectors in a public forum.

If you want, you can hit me up at @Ayjayem on Telegram.

I see how your approach fixes the inequality regarding “interest of interest”, however it doesn’t solve the liquidity advantage for large holders. Large holders simply get liquidity faster due to the 1ICP treshold. Automatic compounding doesn’t change that. Of course 10ICP vs 1kICP get the same reward speaking in terms of maturity via compounding, however large holders still reach the 1ICP treshold faster due to their high stake.

If there is serious market action, high stakes have got a liquidity advantage, which small guys don’t have.
E.g.: ICP rises to 300$ => 1kICP probably gets 1-2 ICP a week. At 300$ that would suddenly be 300$ to 600$ on liquidity in a week. The high stake holder can dump it, the small guy (10ICP) can’t (even if it would only be a fraction; we speak about 0.01 to 0.02 ICP a week =>@300$ : 3$ to 6$).
It’s the simple fact that the high stake holder can (and will) use it to realise profit and engage in the market, while the small fish can’t.

This has nothing to do with fairness and leads highly to inequality. Why does someone with a higher stake get that advantage over smaller fish? There is simply no reasonable explanation.

Lowering the treshold, wouldn’t make it fair, it would only make it less unfair.

@wpb Pegging the condition to a fixed maturity would be interesting, as you’ve suggested. I actually can’t think of any scenario on how to get an advantage right now, since each neuron (same age) would get percentual ICP released at the same time. That would be worth a try. However compounding alone won’t help.


@lastmjs
Your townhall proposal is great, and I’ll give your podcasts a try.

@ayjayem Thx for the invitation. I’m looking forward to hearing more about it.

1 Like

@MichaelAnnh you make an excellent point regarding the liquidity advantage of large stakeholders. Thank you for illustrating it so well.

Automatic compounding, whether it is daily or based on a fixed threshold maturity, provides equality for accumulation, but not equality for liquidity. I’m in favor of equality for both accumulation and liquidity, so I would advocate for a fixed threshold maturity to Spawn Neuron instead of the 1 ICP threshold.

I think a fixed threshold maturity of 1 % is reasonable (although I personally would not object to it being higher). At the current voting reward rate of return, a fixed threshold maturity of 1% means neurons with 8 year dissolve delay would be able to Spawn Neuron approx every 12 days and neurons with 6 month dissolve delay would be able to do it every 25 days.

I have no problem with this difference in timeline because it is another mechanism that makes it more attractive for longer term investing, which is the goal of the tokenomics. This would not be an inequality in my mind, this is the investment choice that everyone gets to make and it applies equally to everyone.

All of that said, I still really want to advocate for daily automatic compounding for anyone who chooses accumulation as their preferred participation strategy. My points above are only intended to address the inequality of liquidity between large and small stakeholders. Implementing a fixed maturity threshold to Spawn Neuron does not negate the ability to also allow automatic compounding or automatic daily compounding.


I am definitely not sold on the idea that neurons should be transferrable. I would only be interested in that idea if the Dfinity foundation confirms that they believe there is a security risk that exists by not making them transferrable and they believe there are no mitigation strategy options. Unless that happens, I would rather stick to the current tokenomics design which is intended to prohibit neuron transfer due to security risks and the desire to attract long term investors.

To me, the free market trading that should exist is the ICP token, not the neuron. If an individual decides to participate in governance by staking, then they need to go into that decision with eyes wide open after DYOR. It’s a choice and all investors should understand the commitment and live with the consequences. I don’t even have sympathy for seed investors (etc) on this point because part of their investment risk was a commitment to accept the tokenomics that Dfinity had the right to design in the best interest of the project. I just don’t agree with the idea that neurons need to be transferrable in a free market when they were not designed for that purpose.

3 Likes

@wpb
I’ve seen that a “Merge Maturity” button was included. I’m really astonished and happy about this. This helps to fix an underlying problem.

As you’ve written:

Automatic compounding, whether it is daily or based on a fixed threshold maturity, provides equality for accumulation, but not equality for liquidity. I’m in favor of equality for both accumulation and liquidity, so I would advocate for a fixed threshold maturity to Spawn Neuron instead of the 1 ICP threshold.

The merge maturity button helps to provide equality for accumulation now. I hope the liquidity aspect will also be fixed soon, via your approach.

4 Likes

It looks like this button is always active. Do we know how much it costs(if anything)? Can I mash it like “abba” in Ikari warriors?

2 Likes

hi all, i recently lost some funds using the NNS wallet app. (see NNS Disburse changes account address randomly? NNS doesnt validate account. Lost ICP - #3 by superduper)

Apparently we are able to transfer funds to invalid account addresses: so i’d like to propose for consideration that the NNS (a) return any funds which are sent to invalid addresses (b) enforce checksum on accounts for transfers of ICP so that in the future ICP cannot be sent to account addresses which don’t actually exist.

thanks

2 Likes

i don’t like this proposal it adds another layer of complexity to spawning which just isnt necessary. we already have enough to deal with and the 1 icp threshold is easy and simple to understand.

2 Likes

i don’t see this as a problem. someone’s invested more in this system and they get an advantage.

whales also have the issue that they cannot sell all of their coins immediately or have to find otc buyers, as well as invest in better security and systems for evaluating proposals. so if you want to get into “equivalency” there are may facets of the prism to see reality from besides when a neuron reaches that magical 1 icp.

2 Likes

I don’t see any value to the network in being able to transfer neurons. I see a big security risk for myself and other neuron holders. Also it will create a sort of “votes for sale” situation that will misalign incentives for long term growth that the network benefits from long term staking when some people will decide to sell their neurons or some rental market comes into play where basically it becomes cheaper to attack the system.

anyway if neurons can be transferred then we might as well get rid of staking because what’s the point then. as no one will have anything at stake since they can sell immediately.

3 Likes

@superduper

i don’t see this as a problem. someone’s invested more in this system and they get an advantage.

So you want someone with more money to have an exponential advantage? Where does your assumption come that 10% invested of 2 people with different net worth should lead to an advantage for the one with the higher stake?

Large holders have to find Otc buyers for their 100ks of ICP… That’s indeed such a problem if you have net worth of millions of dollars and you have to knock on the door of an exchange. Indeed, I’m so glad that I’m not a millionair and won’t have to deal with such things. *Sarcasm off

The only reasonable argument is the security aspect, if you have a higher stake. So large holders want to get something extra for having to deal with higher security for their stake? Is this your main concern? Just to add a few points to this aspect:

  1. If yes, it should be discussed and not be hidden behind the shadow of unclear tokenomics.
  2. The exponential function is hard to handle. As I’ve provided in multiple comments, the advantage of this would multiply over time to such an extent, that a large holder could probably buy up a whole country just to secure his stake.
  3. IMO the protocol should aim for as much decentralization as possible. Since that can only happen over governance, it should mean that neurons/votes/holdings should be distributed over time. Additional incentives to prefer larger holdings are a bit contrary to this. @lastmjs made a great podcast about this.
  4. Large holders have already got a big advantage they always seem to forget: Their large holding. Someone with a large holding can live from the staking reward atm. Spending a few ICP on additional security seems to be a fair trade for beeing able to practically not work and earn as much as a few middle class workers a month.
  5. Comparison to other projects: Eth PoS doesn’t give you more just because you have a high stake. Also not Polkadot, or Cardano. So why do we want to go a complete other way here again?

At least we got rid of this with the merge maturity button now, so the only discussion point is the liquidity aspect. One of my arguments:

If there is serious market action, high stakes have got a liquidity advantage, which small guys don’t have.
E.g.: ICP rises to 300$ => 1kICP probably gets 1-2 ICP a week. At 300$ that would suddenly be 300$ to 600$ on liquidity in a week. The high stake holder can dump it, the small guy (10ICP) can’t (even if it would only be a fraction; we speak about 0.01 to 0.02 ICP a week =>@300$ : 3$ to 6$).
It’s the simple fact that the high stake holder can (and will) use it to realise profit and engage in the market, while the small fish can’t.

If you want to say that this is justified because of greater expenses for security, then I’d like to have some calculations/facts and discussions about this.

What is an acceptable amount of money for something like that? Is the 1ICP treshold liquidity advantage good for this? Should we make another incentive? E.g. Dfinity treasury wallet pays for this directly, instead taxing the small holders indirectly via the liquidity over the 1ICP treshold? What do others say to the presented points above?

Why do other projects don’t have this kind of arguments/problems?
Of course they have other approaches via stake pools and nominated proof of stake, however they are not so different. The stake pool operators collect a certain amount of fee of the people, who are delegating to them.
=> If that is translated, it would mean that the neurons we’re following should collect fees. If someone with a high stake wants to collect additional income, he should advertise his neuron and make others follow him. Would that be an approach suitable for Dfinity?
Via this he could earn without giving others a liquidity disadvantage, he could pay for additional security, and it would lead to more decentralization of the IC, since direct participation in governance becomes rewarded more. Of course he’d also have to work for his earned ICP, that’s what I see as an absolute win for all.
Work more => earn more instead of have more => earn more


@superduper
My main concern in the neuron discussion is that the free market will find it’s paths anyway. Surpressing free market processes over the long term, only destabilizes the system. Actually we could both sell our votes atm. It doesn’t make sense but we could. You could call out a price and I could answer. Buying votes without transfering ICP or neurons. It’s possible.

I hope there will be a Townhall soon for all this kind of discussions.

I agree on this and would be interested in the Foundations reasoning behind setting the threshhold to 1 ICP instead of a certain maturity as communicated @diegop

The maturity of a neuron starts at 0 but increases with voting activity. When a neuron’s maturity grows beyond a certain threshold, then it can spawn a new neuron containing newly minted ICP, which resets its own maturity back to zero.

[source: Understanding the Internet Computer’s Network Nervous System, Neurons, and ICP Utility Tokens | by DFINITY | The Internet Computer Review | Medium]

Another thing I would like to suggest is the option to lock a neuron in “merge maturity mode”. This would automatically merge the maturity (maybe once a day) and not allow neuron owners to spawn any neurons to retrieve their voting rewards. The lock-up mechanism would be exactly the same as it already is and users would benefit from an additional bonus, just like the age bonus.

1 Like

For anyone who is interested in discussing these topics face to face, we’re having a Town Hall video call meeting to discuss these governance/tokenomics topics this Friday at 5PM UTC in the DFINITY Devs Discord: DFINITY DEV OFFICIAL

I know it is short notice, the first time is tricky. But if it goes well and seems useful we can iterate on the process and hold these meetings on a regular basis or ad hoc whenever is necessary.

I hope to see you there.

3 Likes

“most of the ICP minted is from staking not node providers”, for now*. We only have ~250 node machines right now. Should be 100x at least to really start flexing off the computational capabilities. And to have any global impact, it should eventually have 1M+ machines. Otherwise the computational fire power is just way to low to host global apps.

1 Like

Nice thing is that node provider capacity can scale with IC computational demand, which itself drives token price - meaning that each ICP would go further in paying node providers, when there’s a need to bring more node providers on to the network.

So simplistically, if IC canister demands grow 2x, suppose we need 2x more subnets (if already at capacity); this costs 2x as much to pay the new nodes on those new subnets.

But it’s reasonable to assume that ICP price grows roughly 2x in the same period (at least - probably more if you buy into Metcalfe’s Law), and so still the same amount of ICP must be minted to pay node providers (who get paid a relatively constant fiat amount, even if remitted in ICP). So roughly the same proportion (or even a falling proportion) of minted ICP should go to node providers vs. stakers over time, by this logic.

1 Like

How about a fork? Looks like part of the community is starting a new project that is a fork of icp called
icpreboot

Fwiw, I agree with your proposal to set a threshold based on neuron maturity, if that’s possible.

It might be helpful if @diegop could provide the team’s rationale for setting it to 1 ICP in the first place.

I agree with being able to transfer neurons, but not instantaneously. I think users should have to express their intent to transfer their neuron by initiating a countdown (weeks or months). During this countdown the neuron should no longer be allowed to vote. once the timer runs out the user should be able to set up a contract whereby another user can deposit a specified amount of ICP. If the correct amount is deposited the neuron ownership will be transferred and the transfer delay immediately reset.

Just my two cents on the matter

2 Likes

I agree in general. But the demand elasticity is much greater than our capabilities to scale up the fleet size. The IC computational demand is already greater than what we have. Even a single problematic data center right now can slow down a whole region. Looking at the pace of II anchor growth (exponential), canister growth (exponential) and node fleet size (linear, or on pause?), what is obvious is that the computational capability does not grow as fast as everything else. We talk about “boundless scalability” and that is a great achievement on the software side but what it eventually comes down to is the computational capacity, which has problems. The interest is there, I’m waiting in the queue myself, but there is no visibility and seemingly nothing is happening. I’m sure a lot of work is happening behind the scenes though.

2 Likes

Social recovery might be the solution to this: Why we need wide adoption of social recovery wallets

3 Likes

We just finished up our first community governance meeting, what I’ve been calling a Town Hall meeting.

Here’s my summary of the most important conclusions and action items. Hopefully others in attendance will share their take-aways if beneficial:

  1. Neuron transfers seems a contentious topic: should we try to restrict them completely, put in some restrictions, or open transfers completely? Seems like we need a lot of deep thoughts and persuasions on this topic
  2. How will the inflation/deflation of the IC play out over the long term? I think we need someone to do a deep dive and create models mapping out various plausible scenarios. We want to make sure inflation will be in check at future theoretical levels of usage. The relationship between cycle burning and node operator reward inflation seems most important to understanding deflationary pressures…or perhaps just cycle burning. We need to know what amount of burn we can expect in the future. Charts and graphs please, if someone is looking for something beneficial to do this is an open opportunity.
3 Likes

I see neuron sales/transfers kinda working like life insurance policies or annuities sales…where you would have to take a hit or get paid out cents on the dollar to cash out early but it could become an interesting market on an exchange. Unique compared to other tokens.

Example: a neuron with 100 ICP with an 8 year delay.

Why would someone sell? They need emergency liquidity maybe…

Would you pay market price for that? Of course not because you could just do that on a normal exchange. So you would buy at a discount and your return would be the market price after dissolve.

I don’t see the vote buying argument. If you wanted to buy votes go buy up tons of ICP on an exchange and stake but you could get votes at a discount but only if people are willing to sell at a loss.

2 Likes