The future of AI lies in understanding causality, not just correlation.
Good means not merely not to do wrong, but rather not to desire to do wrong.