Y’all are over-complicating these AI-risk arguments