What happened to the whole 'Occam's Razor' line of thinking? Whoooosh. Faster than a Model X, it disappeared into the night!
I don't know--I actually think Occam's Razor works pretty well here, too. IIRC the context of his statement was in response to a question about when they planned to show the Model X in advance of selling them, something they did with both the Model S and the D. His response was basically, "we're not going to do that with the X; the first time people see it is when we're making deliveries." He then went on to talk about some features that would be a surprise.
That's a far cry from a vow to keep the car perfectly hidden and not test prototypes on the road. No offense to any that hold an opposing view, but I think you're reading more into his words than they necessarily bear.
Re: mules--yes, I agree that's what mules are (and what they are for). My point is that, at some point, you've done all you can with a rough, cobbled together mule. You've got the gross calibrations for the powertrain and suspension done, etc. At that point, you have to build prototypes that aren't rough mules--they need to be closer to production spec, so you can test fitment of the various components, and so you can start to do finer calibrations.
All I am saying is that with launch really only weeks away, they absolutely, positively have prototypes extant that are very close to production spec. They may still have some cobbled together mules, too, but the kinds of things you can learn from those are hopefully things you've already learned.