Many applications rely on the system only, so custom chip emulation is not needed for those.
As long as most new applications are just a daydream, the reality is that the dream can only be fulfilled with running legacy 68k apps.
68k emulation right now is very important and possibly forever, regardless how the users feel about it in theory.
While a recompilation would of course help, chances of getting any source for most of the applications are very slim, so the only way to run many things is through full (everything without ppc code) or partial (only some libraries not in ppc) 68k emulation.
You can do this emulation much faster and a hell lot cheaper on any x86 or mac, but if someone has a lot of money to spend on this, and makes his day, let him be happy with it.
I am sure most of us have hobbies or passtimes that we spend on, yet everyone else minus the people walking in the same shoes would just call it a complete waste of money.
The main difference though is that I do not have illusions whatsoever about these activities, and their real worth in the real world - but that's my problem
So like it or not, the 68k emulation will be the case until a new application comes or the whole thing finally bites the dust.