Comment by nopinsight

Extrapolate a bit to when AI is capable of long-term, complex planning, and you see why AI alignment and security are valid concerns, despite the cynicism we often see regarding the topic.