> I don't see how spam or fraud brings value to advertisers It doesn't; that's not the point. The point is, as I said, that the big platforms make money by sel

> I don't see how spam or fraud brings value to advertisers It doesn't; that's not the point. The point is, as I said, that the big platforms make money by selling access to as large a pool of users as possible, and that requires them to not crack down on spammers and fraudsters--because at their scale, it is impossible for them to reliably distinguish the spammers and fraudsters from the rest of their users. So they can't crack down on spammers and fraudsters without also reducing the size of their user pool and thereby killing their business model. The article explicitly talks about this: Either you have to back off and only ban users in cases where you're extremely confident, or you ban all your users after not too long and, as companies like to handle support, appealing means that you'll get a response saying that "your case was carefully reviewed and we have determined that you've violated our policies. This is final", even for cases where any sort of cursory review would cause a reversal of the ban, like when you ban a user for impersonating themselves. And then at FB's scale, it's even worse and you'll ban all of your users even more quickly, so then you back off and we end up with things like 100k minors a day being exposed to "photos of adult genitalia or other sexually abusive content".