Is it your first time using an LLM? No, they generate plausible-sounding bullshit no matter the input. Sometimes that bullshit is useful. Other times it isn't.