Giving this one a separate reply, since it is its own beast.
I’ve thought about this and have made some comments on vote tapering here. I haven’t had time to build a model yet, though.
Such a tapering wouldn’t prevent collusion, though - it would only start curving the power of some NFTs after a certain amount of MNDE locked. Votes from independent wallets would still have to be considered independently, which is why I haven’t spent much time on it: it seems that if a whale wanted to still have full voting power, all they’d have to do is split it among multiple wallets.
I could see a scenario where vote impact curves down not by the wallet but by the accumulated value in the position (Yes/No/Abstain). While this could in theory help against collusion, I currently suspect it would not have had an impact on any of the proposals we’ve passed, where the voting has been mostly one-sided. It could also add some confusion, since users never know exactly how much their vote will count for, and could be demoralized by seeing an even smaller impact if they vote after a whale.
All these complications are why, when it comes to voting models, I prefer to move slowly and deliberately. 