> First of all, let’s be clear that this was all in simulation, something that was not obvious from the tweet making the rounds. The story was obviously a thou

> First of all, let’s be clear that this was all in simulation, something that was not obvious from the tweet making the rounds. The story was obviously a thought experiment from the off before they even came out and said it. And clearly nonsensical. Frankly if a drone AI developed so much real work understanding of what it was, what the remote supervisor was doing. That would be the scary aspect not that it killed someone (that doesn't make sense because no more Yes signals).