Great approach. I also find myself debating Claude on some of my most interesting ideas. It disagrees with me on some issues. I am better than it at crafting good arguments, I think in part because I propose good thought experiments, it almost never does. Instead it runs to cite prior research. But I think the debate may be good practice.
That’s a unique and interesting strength, Jonathan — thought experiments. We tend to forget how individuals can have unique strengths and that those too need to be built into any process of working with AI. Thanks for the reminder!
This is exactly how I wrote academic papers pre-AI. Read enough to be able to make a claim. Formulate the thesis, then read deeper to find evidence to support the thesis. Then write. Like you say, AI is now able to do the deeper read. Without my active participation in the other steps, AI is prone to give me the average of everything out there, which is meh at best.
Exactly! I find too if I don’t bring anything to the table then I get a perfectly competent but boring interpretation. And good connection to how academics often wrote pre-AI. That too was my method but I didn’t make the connection to this piece until after I read your comment. Thanks, Mina!
Great approach. I also find myself debating Claude on some of my most interesting ideas. It disagrees with me on some issues. I am better than it at crafting good arguments, I think in part because I propose good thought experiments, it almost never does. Instead it runs to cite prior research. But I think the debate may be good practice.
That’s a unique and interesting strength, Jonathan — thought experiments. We tend to forget how individuals can have unique strengths and that those too need to be built into any process of working with AI. Thanks for the reminder!
You had me at "Knowing what you want to say, and why you want to say it, is key."
Ah, thanks Rich!
This is exactly how I wrote academic papers pre-AI. Read enough to be able to make a claim. Formulate the thesis, then read deeper to find evidence to support the thesis. Then write. Like you say, AI is now able to do the deeper read. Without my active participation in the other steps, AI is prone to give me the average of everything out there, which is meh at best.
Exactly! I find too if I don’t bring anything to the table then I get a perfectly competent but boring interpretation. And good connection to how academics often wrote pre-AI. That too was my method but I didn’t make the connection to this piece until after I read your comment. Thanks, Mina!