It's a similar worry to AI making an entire world of paperclips, right? If AI systems are produced by the catallactic competition of a lot of parties with their own interests and fed Nostr data (which helps prevent the spread of lies), You could expect to organically have AIs that are farther and farther away from the paperclip dystopia and if one of them becomes somewhat paperclippy, it will eventually be exposed by other models, I guess 🤷
Discussion
yes rouge AI can be fixed by consciously curated AI
Another thing is that balancing games (making them more fair) is by itself an extremely difficult and time-consuming task.
Cutting a broken strategy in half and letting it evolve through the learning of the players is a lot smoother and simpler than trying to eliminate the bad thing inmediately, which could make the state of the game even more broken
That's why "I think" it's sustainable. It's simple and stupid in plain sight
this looks like at least starting the negotiation and keeping the negotiation going in order to reach a deal. both sides want a thing they don't know where the equilibrium is.