It's an advantage, because each user can apply different restrictions to Open Source AI models. If I'm writing a crime novel for adults - I want the assistant to be able to talk about sex and violence. But obviously - a corporate customer support chatbot should not talk about those things.

I understand why ClosedAI added those restrictions. But they are too inflexible.