Rules and models destroy genius and art.

A rule is a record of what already worked for someone else. Learning it makes you competent, and competence is worth having, but a rule cannot tell you the thing that has not been done yet. Work made by matching a model will always sit a little behind the model. The danger is not the rule itself; it is the moment you stop asking whether it still fits what you are making.

Take one rule you follow in your own work and break it deliberately on something small, just to see. Keep the ones that survive the test, and quietly let the others go.