Earlier quoted context omitted.
What are their other options? Imagine they could, with great accuracy, predict whether somebody is going to leave a good comment or a bad one. Do they expose that to the commenter, saying they've been blocked from commenting because the algorithm thinks they're a shitty person? Imagine the optics of that... The other option is the shadowban, where toxic commenters are hidden without notice. It sounds like a good idea…
If the structure of the service offering exceeds the limits of available technology, then the structure of the service is at fault and needs to change -- wishing away the bad people doesn't work, nor does pretending the issue does not exist (their current strategy) until advertisers are forced to threaten them The option available to them is clear, it's just not something anyone is willing to accept: connecting untru…
>The option available to them is clear, it's just not something anyone is willing to accept: connecting untrustworthy anonymous third parties to the bedrooms of 5 year olds cannot be done safely within the limits of existing technology.
Would you ban beaches as well, because creeps might watch others on a beach?