Every user has their own view of what should or should not be abusive. HW's scope is clear and generally applied equally to each case. Of course, I take into account many different factors for blacklisting of downvoting, such as whether the user is new, intentionality, frequency of abuse, etc.
I do not blacklist new users. Neither do I downvote their new posts with abuse. New accounts that are downvoted are those that were evidently created as made-to-abuse.
If I see a new user who posted some AI spam or plagiarism, I would prefer to drop a general comment with guidelines, not specifying exactly what the abuse is. That's a different approach that I had in the past.
I am glad to see your considerations have evolved over time. It is difficult for users to grasp the need to provide original editorial content with material from other creators to qualify for rewards, as well as attribution for memes in particular. No other platform has such requirements for memes that I have seen, and they are simply reposted freely without sourcing.