I've answered this before, but the gist is that we've looked very closely at it and understand both the theory and the math. I will say we have zero plans to give up on proper component-level modeling, however.
The thing with NAM is that it's not particularly scalable. Other methodologies like what're found in IK and QC tend to work better in embedded, non-desktop environments.
It wasn't so much Kemper's patent that's kept Line 6 from pursuing profiling as us having respect for Christoph. And not wanting to be copycat douchebags.
Interesting word choices.
So you're saying maybe ... something like a capture. :-)