Does TikTok have to know that “as a category blackout videos are bad” or that “this specific video is bad”.

Does TikTok have preempt this category of videos in the future or simply respond promptly when notified such a video is posted to their system?

Are you asking about the law, or are you asking our opinion?

Do you think its reasonable for social media to send videos to people without considering how harmful they are?

Do you even think its reasonable for search engine to respond to a specific request for this information?

Personally, I wouldn't want search engines censoring results for things explicitly searched for, but I'd still expect that social media should be responsible for harmful content they push onto users who never asked for it in the first place. Push vs Pull is an important distinction that should be considered.

That IS the distinction at play here.

Did some hands come out of the screen, pull a rope out then choke someone? Platforms shouldn’t be held responsible when 1 out of a million users wins a Darwin award.

I think it's a very different conversation when you're talking about social media sites pushing content they know is harmful onto people who they know are literal children.

The ingenuity of kids to believe and be easily influenced by what they see online had a big role in this ruling, disregarding that is a huge disservice to a productive discussion.

What constitutes "censoring all of them"

Trying to define "all" is an impossibility; but, by virtue of having taken no action whatsoever, answering that question is irrelevant in the context of this particular judgment: Tiktok took no action, so the definition of "all" is irrelevant. See also for example: https://news.ycombinator.com/item?id=41393921

In general, judges will be ultimately responsible for evaluating whether "any", "sufficient", "appropriate", etc. actions were taken in each future case judgement they make. As with all things legalese, it's impossible to define with certainty a specific degree of action that is the uniform boundary of acceptable; but, as evident here, "none" is no longer permissible in that set.

(I am not your lawyer, this is not legal advice.)

Any good will attempt at censoring would have been as a reasonable defense even if technically they don't censor 100% of them, such as blocking videos with the word "blackout" on their title or manually approving videos with such thing, but they did nothing instead.

> TikTok is in trouble for not censoring them after knowing they were causing harm.

This has interesting higher-order effects on free speech. Let's apply the same ruling to vaccine misinformation, or the ability to organize protests on social media (which opponents will probably call riots if there are any injuries)

Uh yeah, the court of appeals has reached an interesting decision.

But I mean what do you expect from a group of judges that themselves have written they're moving away from precedent?

I don't doubt the same court relishes the thought of deciding what "harm" is on a case-by-case basis. The continued politicization of the courts will not end well for a society that nominally believes in the rule of law. Some quarters have been agitating for removing §230 safe harbor protections (or repealing it entirely), and the courts have delivered.