tomrippity02 :
Unless by "different thing" you are talking about the software being so different that any beta applications simply don't work, which I doubt will happen.
Between seemingly neck-breaking unnatural head gestures (I hope the "reviews" I have seen were exaggerating), indiscreet voice commands (everything has to begin with "ok glass" which people can use as a cue to shout their own command if they want to annoy you since the voice recognition software does not appear to be able to tell voices apart) and "point-and-blink" input methods, which one would you prefer?
I am pretty much certain I would prefer eye-tracking for speed, accuracy, convenience, versatility, intuitiveness, discretion/privacty and just about all other metrics I can think of for primary input except for dictation (assuming I have a relatively quiet and private space to use that in) and maybe augmented reality navigation to some extent where head gestures would likely get used for coarse navigation and point-and-blink for details. I imagine eye-tracking would supersede all other Glass input methods almost immediately as the preferred input method - it is the most natural method I can think of for hands-free interaction with a HMD-GUI.
So, what I expect to happen is early developers will spend tons of time, effort and money developing head gestures and voice commands for Glass Explorer out of necessity since the prototypes lack the necessary camera but after Glass and equivalents come out with eye-tracking, most people will end up flocking to eye-tracking for primary input and voice commands or gestures for device wake-up and application launch shortcuts out of necessity to conserve battery power - keep the camera and image processor powered down between uses since image processing may have become cheaper but it is still nowhere near cheap enough to do continuously from standby.
Sure, software for the old Glass will most likely work with "New Glass" but how long do you think "New Glass" owners will remain interested in "Old Glass" antics? Ergo, "Old Glass" and its applications getting superseded and most of that development time going down the drain unless I am wrong about eye-tracking being vastly superior in most cases. (Apart from the energy cost of course.)