ever notice that, with LLMs, the main goal of the people who own them is to turn them from machines that spit out information they gathered in realistic-ishly human ways to machines that spit out what their owners want people to believe?