Content pfp
Content
@
0 reply
0 recast
0 reaction
Varun Srinivasan pfp
Varun Srinivasan
@v
Should we make Warpcast's spam categorization publicly available? Image a 5 category system for every account: not spam, maybe not spam, unknown, maybe spam, definitely spam. What would you do with this if it was available today?
19 replies
17 recasts
154 reactions
Varun Srinivasan pfp
Varun Srinivasan
@v
The main hesitations here is that spammers can use real-time rankings to find ways to avoid labels. But if its useful to people we should make it public.
3 replies
8 recasts
51 reactions
Reeeny pfp
Reeeny
@reeeny
Notify users that they are considered spam along with guidelines on what to do to not be labeled spam. If they are bots, there is no harm done and if they are real people they will appreciate the heads up - if the message is well written.
1 reply
0 recast
1 reaction
kompreni ๐Ÿ…ฑ๏ธ โšก pfp
kompreni ๐Ÿ…ฑ๏ธ โšก
@kompreni
Yes but can u do it on the dnd spectrum?
0 reply
0 recast
1 reaction
Michael Pfister pfp
Michael Pfister
@pfista
Weโ€™d use it to improve @buoy, making the tiers available as search filters
1 reply
0 recast
8 reactions
Deployer pfp
Deployer
@deployer
would probably look into using it for /lp channel moderation.
0 reply
0 recast
2 reactions
Linda Xie pfp
Linda Xie
@linda
Yes we would use that for Bountycaster in filtering out bounty replies and bounty posts
0 reply
0 recast
0 reaction
Garrett pfp
Garrett
@garrett
yes would be interesting for people to experiment with the different filtering
0 reply
0 recast
0 reaction
jacopo pfp
jacopo
@jacopo
could be especially useful if the list was onchain and periodically maintained like neynarโ€™s user score would add support for it on @slice
0 reply
0 recast
0 reaction
Royal pfp
Royal
@royalaid.eth
its actually a pretty difficult question to answer because people could start to reverse engineer the spam filtering from the labels and train around them. It would be a net boon to the ecosystem in the short term but it might be a PITA long term. FWIW I don't feel strongly one way or the other
0 reply
0 recast
0 reaction
Chinmay ๐Ÿ•น๏ธ๐Ÿฟ pfp
Chinmay ๐Ÿ•น๏ธ๐Ÿฟ
@chinmay.eth
I would use it to set message send limits at @percs
0 reply
0 recast
1 reaction
jd ๐ŸŒบ pfp
jd ๐ŸŒบ
@jdlewin.eth
open source is always the way sailor
0 reply
0 recast
1 reaction
Chinmay ๐Ÿ•น๏ธ๐Ÿฟ pfp
Chinmay ๐Ÿ•น๏ธ๐Ÿฟ
@chinmay.eth
If the models are open, the cat and mouse game accelerates. Do you have resources to play the game?
0 reply
0 recast
0 reaction
Colin Charles pfp
Colin Charles
@bytebot
Personally, I would find ways to mark as not spam. There should be some kinda way to ensure the algorithm doesn't negatively penalise real accounts! Otherwise it ends up a little like the X shadowban
0 reply
0 recast
1 reaction
base pfp
base
@basewtf
I would be interested, Im working on a bot/ai agent and I think in the process of testing I may have gotten it marked as spam...maybe my account too?
0 reply
0 recast
0 reaction
frederick pfp
frederick
@sgniwder
i'm leaning against this, but admit i could be swayed. while i believe that it is good to share this on an open platform, I'm afraid that it would be used by bad actors...
0 reply
0 recast
0 reaction
Breno pfp
Breno
@br3no.eth
Definitely! This would help me notice if I'm being unwillingly spammy, and adjust how I cast
0 reply
0 recast
0 reaction
Pichi ๐ŸŸช๐Ÿ–๐Ÿน๐ŸŽฉ ๐Ÿก๐ŸŒธ pfp
Pichi ๐ŸŸช๐Ÿ–๐Ÿน๐ŸŽฉ ๐Ÿก๐ŸŒธ
@pichi
Absolutely.
0 reply
0 recast
0 reaction
Sayonara pfp
Sayonara
@sayo
Make the feed algo open source??
0 reply
0 recast
0 reaction
frdysk.framedl.eth pfp
frdysk.framedl.eth
@frdysk
would be the deciding factor if I should reply or not
0 reply
0 recast
0 reaction