When you run into something you don't understand in a system, what do you assume?
That it's dumb or that you just don't understand why it was put there yet?
If you assume it's there for a good reason, you can learn from it.
If you assume it's there for a bad reason, all you can do is figure out how to route around it.
Things that were put there by humans in the past were likely put there for a good reason, even if it's not obvious.
A key asymmetry: people only tend to do it when they think it will create value in some way.
This is the insight behind Chesterton's fence.