I write a lot of what I call “prayers” — long prompts the are unlikely to work at all. I assume the thing is as smart as a human and then study its mistakes. I then search for variant instructions that improve reliability.

→ View original post on X — @goodside