Was thinking about this the other day, and it inspired me to start working on what I call VibeTag. The idea behind it is to be able to tag any content on the internet with a “vibe”/“sentiment”, whether it’s a nostr post, something on Mastodon, Twitter, even entire websites. This will initially focus on whether the sentiment of the content is positive (informative, funny, etc) or negative (misleading, rage-bait, vulgar in general) etc. As more and more content is tagged in either direction, it learns what words/phrases are associated with positive content vs negative. The tags and info about associated content would be synced across content through the beauty of Nostr relays. This essentially allows there to be a straightforward algorithm that you can take with you across the internet that automatically generates/updates a list of words associated with shitty content and hides it from view. This is something I’ll be integrating into Dirty Gull but also will be a library anyone can integrate into their project!