The point of the ARM laptop would not be to replace your desk-based 'thick-client' laptop, it will be to provide 80% of it's benefits,
No, the
ultimate goal of the ARM chip needs to be to provide 110% of the benefits of your 'thick-client' laptop to
most people. Otherwise, Apple might as well just add mouse/trackpad support to the iPad. Short-term, yes, the 12" MB and the Air are the 'low hanging fruit' that could be replaced by the ARM but long term it doesn't make sense for Apple to effectively have 3 'families' of personal computers - iPad, ARM Mac and Intel Mac.
As has been discussed already, the big, probably unavoidable, loss is going to be for the diminishing pool of people who
need x86 versions of Windows or Linux. The question is, is supporting virtualised versions of
other operating systems that still a sensible priority for a platform who's Unique Selling Point is that it runs MacOS? It's 2019 - a 'real' Windows/desktop Linux machine is the size of a tablet. A virtual linux
server just has to be somewhere on the Internet, and 90% of Linux server-side/web development doesn't care if its running on ARM, x86 or, for that matter, MacOS/Darwin. Then, on the other side of the equation,
everybody developing for iOS and Android is currently working on x86 but targeting ARM (last I looked, the 'emulator' for iOS worked by compiling the App for x86 - so much for completely mirroring the target environment).
As for current MacOS x86 Apps getting 'ported' to MacOS/ARM - that may be harder or easier depending on the product ultimately, the
only barrier to that is whether (a) the developer is still active and (b) they see a market for the product... bearing in mind that getting rid of x86-specific stuff in your codebase might be beneficial for future iOS or Windows-for-ARM versions, too. (It would be interesting to know just how much x86-specific code there is in
current MacOS apps - as opposed to the situation in 2005 or earlier). Honestly, the dropping of 32-bit support, Apple's depreciation of OpenGL and apparent bust-up with NVIDIA, the continuing lack of a conventional 'pickup truck' workstation (or a laptop with a decent keyboard!) are just as big - if not bigger - threats to the future of MacOS applications as a switch to ARM.
It might be portability and battery life that set ARM-based MacBook-replacements apart, but dor the pro range, a switch to ARM offers the possibility of
lots more processor cores, and all sorts of on-chip codecs, vector processors etc. that Intel can't currently offer. The reason that, so far, ARM chips have been targeted at phones/mobile/embedded is that even
Intel couldn't compete with x86 in deskops and workstations (see: Itanium and why the current x86-64 architecture is often called AMD64!).
As I've said before - its not about whether you'll be able to switch on day one. Many won't. The question is whether your work
in 3-4 years time will be helped or hindered by using an ARM Mac. If Apple tries to force the change overnight (as they did with the Mac Pro cylinder, the new MacBook Pro designs etc.) then that's obviously a recipe for failure.
It will be interesting to see if either the 12" MacBook or the Mythical Modular Mac Pro get mentioned at WWDC, and if they're going ARM. An Intel MMMP would commit Apple to Intel for the foreseeable future. If they
don't at least preview the new MMMP at WWDC then anybody still blindly loyal enough to be waiting for a replacement for their trusty but ageing cheese-grater would probably be happy with something based around an exciting new 68.5-bit chip based on the Z-80.