You've haven't thought long enough or read much alignment literature if you don't think LLMs already have considerable potential for harm. In 3 years they will

You've haven't thought long enough or read much alignment literature if you don't think LLMs already have considerable potential for harm. In 3 years they will be devastating. Political interests? Wanting an aligned AI isn't a political interest, it's a basic precaution to not fuck up civilization.