Personally I'd like to see constitution based non-linear flagging to make matters better. Here is the simple overview:
We need a clearly defined constitution to decide which content gets flagged. Then we can use some algorithms (I thought this would need AI but a programmer told me that we don't need anything like an AI for the job) and post these spam content in a different tab.

We encourage people to flag and non-linear flags mean when more people flag, the more powerful the flag become. You can flag outside the spam/abuse tab but unless many people join you, your flags would have less impact. This gives a chance for people to flag content that escape the algorithm while making flag wars hard to continue.

We could even have a separate voting power for flagging abuse.

This is just my suggestion.