I was thinking like crowdsourced flagging and a reputation based system where users via reputation could be granted increasingly higher responsibilities as they use the site. People who post negative content would be subjected to negative feedback loops where their content would be progressively demoted, reduced visibility or encounter contribution limits. Moderation logs would become public for transparency purposes. Data could be anonymized and provided via API for research purposes.

How is this different from Reddit, besides the moderation logs being public? Maybe I'm missing something here, but it sounds exactly the same.

Reddit is heavily centralized in the hands of a small number of mods. This would decentralize things

I'm not sure this holds just by virtue of the system. For example Wikipedia uses a similar model and they are also heavily centralized, the majority of the moderation is done by a very small number of users. If anything, I would say systems like this naturally tend towards a small inner circle.

I want something like you described up until you mentioned moderation. There wouldn’t be a need for it. That’s the whole issue with a single decider getting to suppress content that the rest of your solution solved.

I was using Moderation logs as an abstract concept of the actions taken to promote or demote content. It would be decentralized

I think a better example of this type of moderation would be Wikipedia.

I have used Wikipedia as an editor and the moderation can be atrocious and again in the handful of a few.

> Data could be anonymized

I fucking doubt it.