>In a company that knows more about you than you can possibly imagine, and more about automated sentiment analysis than anyone in the world, they couldn’t algor

>In a company that knows more about you than you can possibly imagine, and more about automated sentiment analysis than anyone in the world, they couldn’t algorithmically determine who should be allowed to post on certain subsets of videos, or devise a system they thought was worth deploying to ensure comments meet a basic level of decency. Those solutions are not perfect and every time Google implemented them they caused collateral damage or even major changes to the whole platform (see the demonetization situation). I believe that they refrain from using them unless pressured to because fixing a problem that is not know to the general public isn't worth the negative PR, complaints, investigations and the damage to stability of the platform.