Re: Tool For All??
I dare say even most people at Google know that user-facing applications don't represent a majority of computing resources used. They're dwarfed by scientific HPC, business backend processing, communications, embedded controllers, and pointless waste like cryptocoin mining.
I gave an academic presentation a few times on "how computers are really used" where I pointed this out, and also noted that it's likely the most common user-facing computer application is "digital clock", because so many appliances have one. (Yes, you can certainly implement a digital clock with non-Turing-Complete hardware, but since those applications these days pretty much always have general-purpose CPUs for other functions, it's cheaper to use them to implement the clock as well.)
In short (I know, too late): While the claims of the Flutterers are certainly hyperbolic, I expect they assumed their audience would infer a "for the user-facing portion of applications that have a UI".
That said, I too find this sort of thing neither novel nor particularly interesting. UI development shouldn't be expensive. UI design is, if it's done right; but these use-it-everywhere frameworks don't help with that. Do the design properly and build the backend properly, and you should be able to cheaply create and maintain UIs for whatever end-user devices you need. They're optimizing the wrong part of the problem space.