@tszzl While it’s extraordinarily useful to understand ai as human like in various properties (“don’t micromanage agents etc”)
we should be really worried about humans extending emotional and empathic response to ai due to treating them as human like (“don’t let me die, etc).