[aobdev]

They’re talking specifically about client-side hashing. You’re right that the system will hash the stolen hash, but only in cases where the system can’t be bypassed (which client side code always can be).