I've never seen anything in print from Mazda indicating that the NA head was redesigned. The write-up in the TSB does say that this new part number represents a loaded head (i.e. including installed valvetrain components), however there is nothing in there about the actual head itself.
Although it might naturally be presumed that the NA head has been redesigned, there's been no proof of that so far, that I'm aware of. And for anyone who insists that the head must have been modified in some way, I've seen multiple references made to 'a batch of bad castings', as being the root cause of the NA cracked heads. I certainly don't know if that's true or not, but if it is then there would presumably have been no need for Mazda to redesign that head.
Regardless of what the actual issue with the head is, this TSB is very unsettling and concerning to me, for one major reason. Why would an automaker take this unusual step of distributing loaded heads to their dealer service departments? The only logical reason I can come up with is that the failure rate must be significant enough to justify that action. If NA cylinder head failures were rare (as is typical with most other newer model vehicles), then why would they ever even have considered doing something like this?
I believe this loaded head enabled Mazda to reduce the number of hours reimbursed to the dealers for warranty head replacements, due to the tech no longer needing to do the time-consuming job of transferring all of the valvetrain components from the damaged head to the new one. And it also eliminates potential mistakes made by the tech doing that transfer work. Just my own theory about that and, like most of this stuff, I'm not expecting to ever have it either confirmed or disproved.