I'm quite happy with the FPGA path.
On the off chance anyone wanted to resurrect the boxer, that would probably be doable without reaching $100 million+ in development.
It would be an upgraded big box Amiga which you could add 24-bit gfx.
Have it also in Mini-itx for a wedge system.
I never got to play with an Amithlon machine, but I assume it would be like WinUAE only better.
I should do this poll in another year. This was more of a quiz than a serious poll.
P.S. I should have added the GBA 1000 option, where you redo the ageing hardware chip for chip.
I like the option of new boards / components the best. Fix some of the inadequacies of the hardware, like faster hard drive controllers (maybe even implement SATA on the board), have built in EEPROMS so we could have upgrades like modern BIOS/UEFI systems, where we can upgrade the libraries in rom instead of having to dink around with setpatch and loadmodules.
To me the Amiga is hardware and operating system. If it were just the operating system, I would use AROS. But if I can replace AROS on an M68k (which looks like it's either getting there, or is already there?) then I probably would! I've always loved the 68k processors. And AmigaOS really takes advantage of it, but I think the closed source nature of AmigaOS is part of why, after Commodore died off, that it didn't remain a very viable option for most people, and quickly turned into a hobby system.
With AROS being ported to 68k, we could replace the closed source system with an open sourced one and finally integrate a lot of the patches that people have been working on, like PeterK's icon.library.
So I would vote for something like Natami (which I think had the best goals in mind) and get AROS to support it 100%.
slaapliedje