Comment by dtdynasty
6 hours ago
I've worked on anti abuse systems and people always say just ban spammers. Just is doing a lot of heavy lifting. People feel the low percentage of spam that gets through is already too much. And when the classification messed up on the low percentage of real users that's also a terrible experience.
The precision and recall tradeoffs are difficult and all scenarios lead to people being unhappy. Not to mention the politics of big corporations where anti abuse systems usually stand in opposition to company goals which makes getting any change through difficult.
Anecdotally, the amount of spam I see on Instagram (the meta platform I’m on) is fairly low. I’m also not suggesting that it’s a a simple problem to solve, but that Meta has deep pockets to pay smart engineers and human moderators to solve it. It just seems unrelated to this change.