Moderation doesn't scale well, so hundreds of engineers don't go very far even if you put them all on moderation. Google doesn't like to put money on anything that can't be automated (even when it's the appropriate, like customer service), hence you see only ML filters that of course can't tell the difference.

Also, careful what you wish for. As we've seen, one man's moderation is another man's 1A violation. I'd much more prefer comments, downvotes and user defined filters to do the accountability, than a guaranteed-shitty DOA moderation/truth keeper team. I worry much more about YT's slow creep towards a fully auto-play, advertiser friendly TV equivalent.