Ask HN: How do I learn about AI safety?

Is there a talk, discussion, book, course or paper that explains what does AI safety mean and then (hopefully) goes on to explain ways to achieve it with the trade-offs involved ?