The problem is that someone (Parler in this case) would have to actively build automated moderation systems or be philosophically aligned with moderation in order to fulfill the TOS in this case since its the users driving the violation. The baseline shouldn't require them build something, that seems like a burden too far. Suppose the violators were bots, doesn't this mean that anyone that can't build a moderation system to keep up with the bots becomes collateral damage?