One big plus is that you need far less acid, because if you process fingers with the PCB, the volume needed to cover them is much bigger. Then again, the acid can attack from full backside of the contact (surface= 1.5 x 4 mm @ PCI) instead of attacking from the 4mm x 8 micron from sideways, which will make AP work within minutes. And furthermore if you soak fingers in AP and the PCBs are multilayers - which most are - AP will also attack the inner layers since they are exposed at the cutting or breaking line, this again will slow down the process and consume more acid and contaminate, since inner layer are sometimes made of nickel.
This guy has prooven one thing for sure: That there are 2 layers with equal amount of basemetals to deal with: Copper AND Nickel. AP is designed to work best for dissolving copper, maybe we should rethink the recipe here, since nickel does play a major role in the process. It ends the AP live and it slows the process down.
I think the idea is good, if you can find a fast and automatized way to peel or scrap them off, it would be the best of both worlds. I hope it is the correct term in english: a metal planing machine would be the right instrument.
Check it out, and watch your fingers when working with one....
[youtube]http://www.youtube.com/watch?v=YpzSXy67HBc[/youtube]