Your belief about whether AI turns out well is not really a belief about AI but about what humanity will do with the opportunity; the future is a design exercise, not a forecasting exercise, so breaking our way depends on making good collective choices—which is feasible but extremely hard.