As far as MOS or OS4 goes....I'll wait until someone writes a PPC G4/G5 emulator core and then lets me run either of those systems via some sort of OS4 emulator 
There are PowerPC emulators out there. I believe QEMU emulates the full G4 instruction set (can't recall if it supports G5 or not,) but it doesn't emulate any OS4 custom hardware, and anyway QEMU's performance is lackluster in my experience. SheepShaver emulates PPC Macs, but I don't know if it goes up to any of the G4s, or if it emulates things well enough for MorphOS to run.
It's been so long since a real Amiga was made (1994) that really, in reality, there will never be a modern Amiga because too much time has now passed to actually create a genuine Amiga setup on level playing field with OS X/Win7 machines for daily use (let alone a superior system which is what the A1000 was in 1985 to EVERYTHING and in EVERY WAY 
)
Oh, that's a load of crap. OSX and Windows 7 aren't powered by fairy magic imported from the distant future, they're just newer, more solid implementations of old technology. Hell, OSX's kernel is based on CMU's Mach project, which is as old as the Amiga itself!
Amiga-like OSes can be hampered by how much they want to retain compatibility with legacy Amiga software and its API, but so were Windows 95 and Mac OS (in fact, later versions of Mac OS suffered from some of the same problems Amiga updates have to deal with, stemming from both systems' decision to rely on a lack of memory protection that hampered things later when MMUs were common and could have helped stability.) Mac OS solved that issue by ditching the existing system entirely and relegating it to a whole OS instance running in emulation for classic applications, not too different from some Amiga approaches. None of this stuff is magic, it's just one set of workable solutions to a common set of problems.
And no, a modern Amiga attempt probably
isn't going to be the absolute best thing on the market, because it's a whole different market now than it was in 1985. If we focused less on keeping up with the Joneses than simply trying to make something that's good on its own merits, we might have more systems worth using and fewer $3000 paperweights that underperform $500 used Macs.
