The Beneficial Reward Reserve - depletion & refilling


Hi everyone,

I was thinking about the beneficial reward reserve and how to keep it functional long term. According to the WP, the beneficial reward reserve will be 80 million AGI tokens which might seem a lot now but might not be enough long term.

This reserve will be used to support beneficial projects. I think there is a risk the reserve will be depleted and thus not functional in long term which could lead to undesired outcomes. To create an AGI/ASI that is benevolent and stays benevolent in the future we need the reserve to be continuously refilled so that it can support other beneficial projects also many years / decades or centuries from now. As we also don’t know when exactly will the general intelligence emerge we should make sure there is enough beneficial projects at any time - prevent a scenario when the whole reserve is depleted by 2030 but the general intelligence won’t emerge sooner than in 2050.

What I suggest is some sort of “beneficial tax” mechanism built into the smart contract. From every transaction (or from sum of transactions of an agent - eg. on daily basis) there might be a very tiny % or AGI tokens sent directly to the beneficial reward reserve to support future beneficial projects. This percentage could be changed (increased or decreased) by a democratic vote in the future so that it is set optimally…

What are your thoughts about this subject? @bengoertzel ?


I agree with you. Any commercial project must earn money to Finance costs. I created a similar theme here I did not get a clear answer why the management chose a loss-making economic model.


I think the platform itself needs to stay non-profit and decentralized. Otherwise we could create a profit chasing AI and would be no different from those AIs that can potentially emerge from big tech companies. I can maybe imagine collecting some sort of commission at the beginning of the project given this commission can be changed or canceled by a democratic vote of a community. On the other way, money corrupts, so better stay non-profit from the beginning.

SNET itself also plans to sell their own AI services on the platform. This can hopefully earn them enough to be able to support the project in its infant stages.

What I am suggesting is not a fee that would go to SNET’s pocket to do whatever they want to do with it. These collected tokens would just keep the beneficial aspect of the project relevant in long term and it’s use would be decided by the community via mechanism described in the whitepaper.


There is also one more concern, not sure this scenario can or can not happen.

Artificial AI agents in the network (mindcloud), when they are somewhat intelligent, they can own AGI tokens themselves and participate in the democratic votes?

What if they have the 65% majority of votes, can they decide what is beneficial? (their logic can be somewhat different from ours)


Yea, not sure this could be possible either.

There are definitely more options how to increase funds in the beneficial reward reserve. There also can be more ways of how to spend it :slight_smile:

Funding development of new beneficial projects as per the whitepaper.
Or even rewarding agents that are otherwise not considered as beneficial if they do something good. I mean, use it as a dopamine reward to motivate agents to do good stuff :slight_smile: