Varun Srinivasan
@v
AMA - anything you want to know about spam after reading this thread
16 replies
12 recasts
133 reactions
@git
crazy how much things have improved. From a point where every cast I posted received upwards of 100 chinese replies to now, where it's better than X and bluesky. with the knowledge you have today, what would you have done differently when you were building the protocol?
1 reply
0 recast
0 reaction
Varun Srinivasan
@v
biggest mistake was trying to make spam classification binary. would have introduced 3 levels to the classifier early on and trained it on those classifications instead.
0 reply
0 recast
1 reaction