I think the main problem is that speaker IRs only translate the change in tone and not the dynamics of the speaker and how it responds to transients. Once this gets solved we will be 99% to the real thing in my opinion.
I'd say it depends! If you want a cab in the room sound, I agree. But for frfr systems, IR does a great job. If you would mike lets say a Marshall Plexi with a 4X12 cab loaded with Celestion Greenbacks and capture the sound of this cab with a SM57 going on a mixing desk and you would listen to the result through studio monitors, it would definitly sound different than the raw amp and cab sound in the room. Now, if you would choose a Plexi amp in the axe 2, with one of the cabs IR with greenbacks, using a SM57, going direct to the desk and then, to the same monitor, my guess is, eith some eq, you'd get so close to the first set up, no one could tell the difference. And that's the point! All the tones we love and wish to recreate in our living room were recorded tones, polished and sometimes heavily process! Even a simple basic amp and cab set-up like ACDC sounds different going through some mics and mixing desk and layering at the mix! And if you're after that in the room sound, use a real cab and poweramp. But you won't get a lot of tonal versality, because the cab sims in the axe or third party affect the tone a lot! But if Cliff can make those cabs sound even better, I'd be a fool not to be happy about it!!
so there is still enough room for things in the Axe III 
(fwiw, it was a fast attack and release (10-25ms) beating away at that low mid region around 280hz with a fairly wide q. That is the main problem area with heavy guitars & cab flub as well as subs less than 80hz and fuzz > 14khz.
