We all have different lines when it comes to how and when AI should be used. Benjamin Breen writes about some of his experiments with it while using it for historical research, and how it can be a valuable tool.
But the architecture of these models, the data that feeds them and the human training the guides them, all converges on the median. The supposedly “boundary-pushing” ideas it generated were all pretty much what a class of grad students would come up with — high level and well-informed, but predictable.
Although this is not exactly a breakthrough finding, it is quite indicative of the sorts of ways that generative AI can help us do research going forward. It is another, wholly alien pair of “eyes” on a given problem or domain, and that altered perspective can be helpful, even (or especially) when it’s wrong.