Community-based "Subnet Bans"

[Jackalgirl]

Is your feature request related to a problem? Please describe.

It's possible for a subnet to be created that brings no value to the network, such as subnet 28 (the "Tao Accumulation Corporation"). This throws off the emissions balance. It's also plausible that subnets could be created for illegal purposes, such as Ponzi schemes or the generation of AI exploitative/revenge/child pornography.

Describe the solution you'd like

Allow the Senate to vote to "ban" a subnet. A YES vote would result in the subnet's emissions to be set to 0 (and, if possible, the underlying swap pool hard-coded to an exchange rate of 1:1) and the network's emissions redistributed accordingly to the other non-banned subnets.

A longer-term issue would be how to prevent meme subnets/pump-and-dump or ponzi schemes from obtaining enough stake to take control of the Senate.

Describe alternatives you've considered

The other alternative is to return to pre-dTAO days and leave the control of emissions to the hands of knowledgeable persons, who would have to both remain vigilant and be trusted. This is somewhat that, in that it would be a group of knowledgeable persons (the Senate) voting to withhold emissions to a subnet, but they would not be determining the proportion of emissions to non-banned subnets.

Additional context

No response

[icetrust0212]

Is your feature request related to a problem? Please describe.
Due to permissionless approach, some bad actors could launch subnets which are not related to AI development or didn't provide any values to the world.
For example sn28 (LOL)
It's just staking service and miners are able to earn staking rewards without any technical service.
It's literally staking service.
As its reward mechanism, it is attracting more stakes, and the alpha price is going up, and it makes sn28 gets higher emissions as time being.
It steals emission from other AI subnets literally, so will discourage AI development.
If we keep this as it is, then there will be hundreds sn28 subnets in the future, and bittensor might be over.

Describe the solution you'd like
To keep decentralization, but have ability to limit this kind of bad actors, I suggested like this.

• Everyone can launch new subnets without any permission.
• Senate can have capability to ban or limit bad subnets through community vote

I believe this won't break decentralization, but can solve current ethic problem efficiently.

[Jackalgirl]

I'd also recommend that the full contents of the proposal (the reasons for it, the proposal statement, the name of the offending project and whatever information about the owner(s) that is known at the time -- more than just the subnet number) be maintained so that people can better understand why this happened. The OTF could maintain this information in the GitHub, and the hash of the entire proposal (and its outcome) could be the thing that's recorded on the actual blockchain record, so that people can verify that historical GitHub data against the blockchain record later if need be.

[mikyoch]

I disagree with introducing the AMM (Automated Market Maker) strategy to the current ecosystem. I believe it strays too far from the initial purpose of the subnets, as contributors may start to focus—if not shift their attention—towards managing the alpha token price.

While I acknowledge the significant effort and time invested in the development, I suggest reverting the current dTAO update.

[Jackalgirl]

I disagree with introducing the AMM (Automated Market Maker) strategy to the current ecosystem. I believe it strays too far from the initial purpose of the subnets, as contributors may start to focus—if not shift their attention—towards managing the alpha token price.

While I acknowledge the significant effort and time invested in the development, I suggest reverting the current dTAO update.

I get where you are coming from, but please submit a separate "Issue" for this request, as it's not really related to my feature request, which does not involve the reversion of dTAO. Your proposal is likely to get lost here. Thanks!