Recent searches
Search options
@andOlga@girlcock.club Gotcha.
Honestly, the folks on this platform need to get over this quickly or large accounts will leave.
As an example of a possibility: somebody ran a reply to me through a language model (specifics unknown) and it gave that reply a score. I'm not saying we need LLMs getting involved, but something like (which gave a score and hid things behind a CW that were above it) that could be, I'd think, rather easily bolted on - even if it was on the client side.
@andOlga@girlcock.club I do not have the technical expertise to do this, nor am I interested in doing so.
And there are reasons I wish it happened on the platform rather than the client - I earnestly want there to be some kind of signal to others about what's good and bad behavior. That's completely and utterly missing here because the replies are just a giant bucket that nobody's gonna go through (but I will see each drop added).
But I'm confident is has to happen one way or another.
this could be done server side:
reply gets send to your server,
your server (idk how) scores it, decides if it sends you a notification or not
when other servers query for those replies, there can just be an extra field in a reply indicating it's score, then that other server chooses how to handle the score, as it could use it as is, or try to recalculate by itself (maybe following different rules, maybe using your server's score as a hint)
@TechConnectify Yeah, I think it's a platform issue (as in ActivityPub) but one that is expected to be fixed client-side (such as Phanpy which is what I use) and/or instance-side (your Mastodon, Friendica etc. instance) which is where the pushback is. I don't really care for the bare bones way notifications works and how T&S is pretty much left to the end user to figure out for themselves, BlueSky has something similar going on and wouldn't entrust them to make folk to feel safe either.