Size Matters: why universal 'mobile-first' is a bad strategy
medium.com
medium.com
re: #2 - this assumes that the quality of workflow/creation tools have peaked. Part of what drives the growth in tablets as a work device is simply that an office filled with tablets is an order of magnitude cheaper to supply than an office filled with workstations. I'll admit some category of activities (eg programming, architecture, anything precision related) will resist the shift to mobile/tablets harder than others, but many white collar jobs don't actually require desktop grade computing, it just happens to be the case that desktops/laptops were all that was available and that is where the tools were. (even this is being pushed - look at the Binary App for development and Plangrid for architecture)
Re: #3 - Marco argues this point re: HTML5 (and Android) really well. With the hardware accelerated graphics made possible by iOS 7 (and the better hardware coming with the next gen iOS devices), you're going to have to work REALLY hard to create HTML5 apps that "wow" like native ones can.
Re #3: Again, this depends on the type of application you're making. Not all great interfaces need max out the hardware (think Dots or Medium).
The studies you cite will also have the built in bias of comparing a 3 - 5 year old computing paradigm with one that's had a quarter century to mature . . .
As for #2, we'll see where things head. I'm just seeing some refocus on better desktop UX that I hadn't seen in a few years.