Here is my issue with the new touch bar as a general UI element. Things meant for fingers should have tactile feedback unless you're also looking at the same location (like a touch screen) where you can get visual feedback. This offers no tactile response when used, and its layout changes constantly, thus it needs to be looked at to use it. That means you have to shift your vision focus from screen to keyboard repeatedly as you move through apps (or even while in the same app if they make the content dynamic). To me this is bad UI design at a fundamental level.
I can see this being handy as a "slider" style control (though a touch pad edge could be as well). Otherwise though, especially when used as "buttons" my opinion is that it is going to be very tiresome to use extensively and slow general interactions down. If you need this kind of interaction style it should be on a touch screen without forcing the eyes to change focus or monitor changes in peripheral vision.
TLDR There is a reason we touch type. Touch. Exactly what this is missing.
I type a lot. But 99+% of that is A-Z and common symbols. While I normally touch type, I use the function row so infrequently that I don't have the muscle memory to correctly locate the f-keys. The only ones I can do blind are esc, volume up and volume down, and even those I only type rarely. All the rest require me to stop and look at the labels on the keys.
Furthermore, I'm not really using the key shape for those to hit them. I'm mostly just hitting them based on their relative position from the home row. So IMHO at least this wouldn't really affect me - If I end up with one of these laptops I'll still probably be able to hit the few I do infrequently use without looking just fine, and for all the rest I'm looking for labels anyways.
I don't think the "shift your vision" part is a big deal, since it's right next to the screen. I do think there is a disadvantage in not having tactile feedback all the same, and it'll be interesting to see how that plays out. As it stands currently, the presence of tactile feedback for those keys doesn't do me any good, because I don't use them. If the touch bar ends up being useful for me, it'll be a net gain even if the experience of locating the controls is worse.
For people who actually use the function keys, I sympathize and I can totally understand how this is likely to be a huge step backwards. But "developers" does not equate to "uses function keys regularly."
I think the idea here is that it's easier to look down than it is to spend a long period of time holding your arm up to a laptop screen. Either you're moving you're eyes off the screen or lifting your hands up off the keyboard - I'd say there are arguments for and against each, and I look forward to seeing how this solution actually works in use as opposed to the armchair theories we've been seeing.
I can see this being handy as a "slider" style control (though a touch pad edge could be as well). Otherwise though, especially when used as "buttons" my opinion is that it is going to be very tiresome to use extensively and slow general interactions down. If you need this kind of interaction style it should be on a touch screen without forcing the eyes to change focus or monitor changes in peripheral vision.
TLDR There is a reason we touch type. Touch. Exactly what this is missing.