Isn't this why AROS exists?
Starting over again on a legacy-free x86 Amiga-like OS would just be re-inventing the wheel. Give AROS a 68k JIT emulator and UAE integration and you pretty much have everything you described above.
Bar PPC emulation, which is not really required, nor is OS4 compatibility really necessary - how many OS4-only applications are there?
I suppose something like MorphOS's OS4Emu compatibility layer could be written (re-oute all calls to OS4-specific functions to the OS's own equivalent). But writing a PPC emulator for x86 is non-trivial, and since the few Amiga PPC applications that are available are almost all available in a 68k version anyway (which would run faster on a 68K JIT emulator than the PPC version) it's not worth it.
Perhaps Hyperion could work together with the AROS team, and donate their own expertise to the project, along the lines of the MorphOS / AROS co-operation that we have seen in the past.
If Hyperion did go it alone, they couldn't call it 'Amiga OS4' any more, they couldn't claim it as the 'official' next-gen Amiga OS, they couldn't claim it was based on 'The real Amiga Source Code', and so on. Instantly losing 70-80% of their fan-base.
'The Name' and 'Official' status of OS4 is / was one of the most important factors in keeping people from moving to MorphOS / AROS / MacOS X / Linux / whatever. Without that, they'd have to rely solely on the OS's merits.
