What a terrible ending to the article. What Musk is warning against is a system that is capable of accomplishing it's own goals, and if those goals conflict with our own goals then there will be trouble - especially if the AI is clever enough. Whether or not we are living in a simulation is irrelevant - we are still agents with goals.
Maybe the author was trying to be silly, but, if that's the case I can see why he writes for CNet at not Comedy Central.
I think even before AI becomes intelligent enough to potentially destroy us. They will be intelligent enough to get used as smart weapons that are controlled by only a handfull of poeple that seek to controll the world. For Atomic bombs you still need a lot of people to keep them working and effective. AI Robots on the other hand could work, repair and advance themself.
Maybe the author was trying to be silly, but, if that's the case I can see why he writes for CNet at not Comedy Central.