been thinking about fantasy/scifi rule systems and free will
Machine rules are actually the same as fae rules, with the loopholes
It’s why people who entrust a task like “optimize the organization of the database” to an AI will find out that technically the most efficient way to “make a bunch of files not be disorganized anymore” is to delete them all
It’s canonically the reason Asimov’s First Law includes the “and don’t allow harm to a human through inaction” clause. Because otherwise (according to the book) if a robot wanted to harm a human but couldn’t because it was programmed not to, it could always just do something like: drop a heavy object from a great height above the human… tell itself that it was gonna grab the object before it hit the human and no harm done… and then change its plan and passively let the human be crushed by the object already in motion.
Machines are fae, and that’s why my rules-lawyering autistic brain loves them