How do people feel about having some of the open-source LLMs offloading some of the moderation burdens? I understand it sounds rather dystopian, but I also feel if the model is developed transparently, it may offer a solution to the moderation problem.