Regarding moderation, it's true that you can't moderate billions of people with 100% accuracy, but you can discourage them from posting undesirable content in the first place by associating real consequences such a a permanent ban (or a monetary loss, by charging an entry fee to create an account) or make them earn the privilege of posting content (for example, not being able to post links until your account has certain reputation of good behavior).

Discouraging people from posting bad content, and not amplifying the reach of bad content for engagement's sake should go a long way.

Instead, these people were groomed over a period of months or years by way of recommending conspiratorial or outrageous content and it finally blew up.

So not only did Facebook create the problem in the first place, they also had plenty of early warnings about what's been going on, but it's hard to consider an increase in "engagement" (thus revenue) as a "warning" and even harder to act upon it.

This also raises another question regarding the efficiency of our intelligence services if large-scale domestic terrorism was organized all in public on a platform they had privileged access to.