> it will be impossible if the filename changes.

Not impossible, just different and arguably better - comparing hashes is a better tool for finding duplicates.

From a technological standpoint, sure. I'd argue when you're staring down the barrel of 19,234 duplicate file deletions, with names like `image01.jpg`, `image02.jpg` instead of `happy_birthday.jpg`, there's a level of perceptual cognitive trust there that I just can't provide.

^ facts