Eliezer’s article is about people taking scenarios from science fiction about artificial intelligence as evidence of what artificial intelligence is like. In a story like this one, the summary itself is the evidence, and I can’t analyze it and explain it to you in anything shorter than a plot summary. If I could, it would be a bad novel. The purpose of this type of novel, as opposed to a Terminator action-adventure flick, is to explore things that are too complex for us to analyze. Any novel that could be analyzed in the way you’re suggesting would be a bad novel.
Just because that’s the specific focus in the article doesn’t mean that the point is so narrow. Just as it’s incorrect to suppose that a sci fi story gives us a useful picture of how society would be transformed by certain technologies, it’s also a mistake to conclude, for instance, that a story about a bunch of young boys stranded on an island who devolve into barbarism is a useful case study in human nature. The contents of the book never happened, it’s just something someone imagined, and to the extent that the author’s belief that such a thing might happen constitutes evidence, we can do better by looking at what reasons a person would have to believe it in the first place.
Any novel whose experience could be replicated via the process I described would be a bad novel, but what you’d be leaving out would not actually be evidence for the truth of the points the novel is contending.
Eliezer’s article is about people taking scenarios from science fiction about artificial intelligence as evidence of what artificial intelligence is like. In a story like this one, the summary itself is the evidence, and I can’t analyze it and explain it to you in anything shorter than a plot summary. If I could, it would be a bad novel. The purpose of this type of novel, as opposed to a Terminator action-adventure flick, is to explore things that are too complex for us to analyze. Any novel that could be analyzed in the way you’re suggesting would be a bad novel.
Just because that’s the specific focus in the article doesn’t mean that the point is so narrow. Just as it’s incorrect to suppose that a sci fi story gives us a useful picture of how society would be transformed by certain technologies, it’s also a mistake to conclude, for instance, that a story about a bunch of young boys stranded on an island who devolve into barbarism is a useful case study in human nature. The contents of the book never happened, it’s just something someone imagined, and to the extent that the author’s belief that such a thing might happen constitutes evidence, we can do better by looking at what reasons a person would have to believe it in the first place.
Any novel whose experience could be replicated via the process I described would be a bad novel, but what you’d be leaving out would not actually be evidence for the truth of the points the novel is contending.