Aside from regular raising/lowering/pinning/blocking of domains, I believe it would be good if users could specifically rank a result as a good/bad response to that specific query.
This could not only help improve Kagi's current search result surfacing, but could also provide better quality secondary data to back up things like SlopStop reports, as if a site is both ranked as having low quality responses, and some people report it as AI-generated, it's easier to just mark it as AI slop and move on vs only having the reports from people confident it's AI-generated.
This could also help Kagi build a better independent search index at some point after enough data from a wide swath of contributors is stored over time.
Obviously this could have some privacy implications due to queries needing to be stored in some form, so I feel like this would probably be something that should be made clear to users, and opt-in only.
It also has similar problems to this post regarding scale and bias, so I think this would be more of a long-term effort to build a better index, with the short-term benefit of corroborating SlopStop reports.
This would be fairly simple on the user's end, just a + or - button next to each result, that can be clicked to signal if they think the result was good or bad, compared to their search query. This type of mechanism already exists on platforms like Reddit with upvotes and downvotes, and often means higher quality/relevant content is surfaced over simple keyword searches.
This is also present in essentially every algorithmic social media platform, though they mostly just have likes, with dislikes being more of an antipattern than an actual button users can press. This system could work well for Kagi too, in my opinion.
