Show Indicator for AI-Generated Tags when Saving
m
mollybee
When I save a bookmark via the browser extension, I now see AI-generated tags mixed in with my own tags. There's no differentiation between them. This makes it difficult to tell which is which, especially since they can be quite similar. Several of them are compounds of tags I created. For example, I just saved a stock photo of a texture, and the app offered up "stock images," "stock photos," "stock," "texture," "images," "photos," "free," "free stock," and only 2 of my own tags, "photos," and "free stuff." Not helpful! If I don't remember the
exact
wording of the tag I created intentionally, I either have to open Raindrop to check, or I have to guess and hope I didn't just create brand new tags for a single item. When that happens, I have to open old
Raindrop to merge the tags. It's confusing, messy, distracting, and frustrating.There should be a clear visual distinction between tags that are already in my library and ones that aren't. Recommended existing tags should also show a count of how many there are in total. Tagging only works when the system is limited and consistent. Having too many tags Personally, I prefer to see how I'm using my tags while I'm using them. It helps avoid excess time spent maintaining my system.
If showing an indicator isn't possible, could we at least turn off the AI-generated new tags? Every time I save something now, I'm anxious about creating more chaos than my own brain already has.
Log In
s
sjenkins95
I'm seconding this, it's tragic to know a really smart feature was implemented and then deliberately ruined. Really, it just needs two separate modes, because there are two camps of people. But I personally can't fathom throwing tons of spammy, random tags into my system. Thats what I used to do before I understood how tagging works.
m
mollybee
sjenkins95 Two separate modes sounds like a great idea, although I think the AI tags would still need to have some sort of indicator. Even if someone wants a bazillion tags with every SEO keyword from the page, I doubt they’d be pleased to find their library full of inconsistently-applied singular/plural repeats, functionally identical compounds, reformats of their own tags, or words clearly unrelated to the purpose of the page.
Existing, obviously relevant tags are so rarely suggested; it makes sense to assume current choices will later be ignored in favor of something less relevant or meaningful. Without even a 👍/👎 button, you just to hope it understands your intention and pieces that together with other interactions you can’t identify. Surely, when people say they don’t want to train the AI, it’s with the implicit expectation that the AI won’t make rookie mistakes on the regular
and
not allow any feedback.