Imagine surgeons who can't touch a surface with his/her hands due to hygiene concerns using it to manipulate various images/scans of the patient on the table in front of him/her.
Well yes. If multiple solutions gain traction, it just proves that there is a market for this kind of thing. It's not about being first, it's about executing well. If Leap is more accurate for tracking movement, it could prove more suitable than the kinect.
The problem is one of digital vs analogue, and ballistic vs servo. On a keyboard, you either press or do not press a key, and assuming that your finger is aimed somewhere in the right box, you need no visual feedback to see where it is and which thing you're touching. You also get the tactile feedback of "yes I hit the key", "I hit the key but it was at the edge, better recalibrate", etc.
So while there are a basically continuous range of positions you can put your fingers into (and with fifty or so degrees of freedom), the differences between many similar positions are subtle and require feedback for you to see which of them you're in---looping in your visual system and slowing down the interaction considerably. Which is what you want for certain sorts of continuous-ish interactions, and not at all what you want for certain sorts of digital-ish interactions.
All of which is to say, this sounds suuuuuper cool, but it's not going to replace the keyboard.