Finding papers that disagree with the one I'm citing

I almost didn't write this down, then I did it again a week later and forgot half of it.

Check the output against one real example before you trust it. One check beats a hundred promises. Elicit helps here because it pulls claims out of papers and shows you the source.

Step three is the part that makes it yours: rewrite the first and last line by hand. Those two lines are where it stops sounding like everyone else.

Step two, give it one job, not five. The moment you pile on asks, the output splits its attention and you get a polite nothing.

I broke it once by asking for everything at once. Narrowing the ask fixed it instantly.

I'd have kept doing it the hard way if someone hadn't shown me this. Now I can't unsee it.