Existential Threats & Alignment Risks from Artificial General Intelligence (AGI) - Eliezer Yudkowsky {LessWrong}