First of all, this is a very good initiative and demand of the most steemit users. Following is just my idea, not an implementation plan:

Trending algorithm can be improved by implementing different upvote options other than reward/flag such as like, dislike, controversial, innovative etc. Also, filters can be used to find trending posts based on those parameters (vote options), similar to reddit's controversial such as the most liked, disliked, controversial, funny. A single criteria may not suffice all the needs of the users.

A blockchain solution can be possible for the above idea. For example, a fixed 1 cent vote for like/great/happy/funny, negative 0.05 cent for dislike/anger/sad, 1 cent for controversial/neutral. A user will have 10/20/30 (perhaps assigned to him/her based on reputation) such votes per daily regardless of SP they hold.

People sometimes do not vote even if they like the content due to limit of their vote power (VP) or already exhausted VP. But with these extra VP, posts can be discovered in numerous algorithmic ways using ML/AI, such as most liked, disliked, funny or controversial post.

I am not sure how other than blockchain similar voting options without any reward that can be implemented.

Kudos to future developers :)