> it's pretty easy to detect like 90% of bot users
How to tell someone’s never worked on anti-abuse at scale
How to tell someone’s never worked on anti-abuse at scale
The problem with anti-abuse at scale isn't catching the 90-99%. It is the remaining 1%. Have 10M abuse accounts with a 99% success rate and you are left with 100k spam accounts. That, and making sure you aren't catching too many innocent customers in your efforts.
I take it as a given that Facebook, NSA, whomever, have uniquely identified every person, living and dead. I just don't understand how or why fake accounts are created.
Smaller outfits like metafilter need some kind of gatekeeper. Like require nominal fee or scan of ID or verification photo.