As a certified bullshitter myself, I often find myself really annoyed with llms because their bullshitting is just so obvious

I would think because ai is basically just a yes man they can get instant gratification from. Easier to manipulate than a real human, when they're wrong, you can berate them without year of pushback.