And now they're usually pretty good, under most of the popular videos it's just good-hearted jokes. There can be a factual correction, but it's rarely a shithole that you can see on other social networks (e.g. Twitter).
There certainly weren't any active moderation on scale, so it must be their design, incentives and ranking algorithms. So, the question is, how they did it?
Youtube comments are a minor feature relative to the major content of videos and streams. They are easier to change via moderation...and there's not a major loss if authors turn them off or heavily screen them.
In addition, Twitter comments are viral via retweets. Youtube comments are unlinked. There's no on-platform network effect to spread "censorship!" outrage. It's all local.
Maybe that was the plan: creators could be more inclined to moderate comments shortly after the upload, and trolls will not try to get their fix if they know nobody will see their comments.