I think Gwern was too focused on academic PDFs posted publicly to support some argument. There, it's actually advantageous to make a completely new PDF with you

I think Gwern was too focused on academic PDFs posted publicly to support some argument. There, it's actually advantageous to make a completely new PDF with your faked data, since it prevents people from finding the original and doing a comparison. But for official documents used as legal proof, staying as close as possible to the original is required for a forgery to avoid sticking out simply for having the wrong formatting. But you usually won't find those posted publicly on the internet.