By the way, if you watch the video and get sucked in by what they are saying, there's a perfectly good reason for the phenomenon they show. If you scan a doc with Optical Character Recognition (OCR) turned on, Adobe Acrobat is capable of taking a scanned doc and breaking it out into layers automatically. That’s what seems to have happened here. There are still one or two funky anomalies, like the text in box 17a getting split into two separate layers and the date in 20 getting its own, but OCR is an imperfect beast and stuff like that happens.
Not that facts like this matter.