Isn't that kind of the point though? We don't need 100 years of video uploads every day. We need fewer, higher quality videos. If 90% of those were likely to get rejected, such as all the low-effort AI spam videos generated today, then there wouldn't be much incentive to submit them in the first place. The volume of uploads would drop. This is only a problem if you're an advertising platform who needs an endless stream of new crap to recommend to your users to try to keep them on the platform as long as possible.

This is why I keep landing on moderation-before-inclusion as the standard. Review systems that flag bad content means that bad content still gets seen, and often still has enough of an effect to make it worthwhile for bad actors to churn out. IMO that's part of explanation of why there's 100 years of videos uploaded to YouTube every day. Even videos that get taken down get some exposure; increasing the volume of crap is a solution for their short life span. It seems like a self-reinforcing problem there, even if there were an honest attempt at moderation.

The old content doesn't go away, but YouTube is incredibly bad at making it effectively searchable. They don't want you to find the highest quality videos on a topic, they want you to plug in to the endless new stream of slop which just recapitulates things which were done better before.

This isn't really an "old man yelling at kids to get off his lawn" take that everything older is better. Instead I'm saying that the Wikipedia approach to help users find and collaborate on good content to make it even better or more comprehensive is preferable to the approach that the only way to be seen is to endlessly publish new content, which is meant to obscure what came before it on a topic. It's like if you walked into a library, but the only thing visible was a "new releases" shelf and searching the stacks was nearly impossible. That would make it basically a matter of luck if you found anything relevant, and makes deep research impossible.