Do you not think Mr Kemper thinks the Axe is inferior to the KPA? Not defending any of Fractal's comments about the Kemper, but your post is a bit misleading. Fractal has a very long history of product refinement. The Tone Matching feature on the surface does seem inspired by the Kemper, but if you dig a bit deeper you will see folks had been EQ matching recorded tracks using Ozone and creating matching IRs to load into the AxeFX for quite some time before the KPA was released, they simply automated that process. It is very different from KPA Profiling which requires an Amp. There was the addition of something similar to the KPA pick attack which was very likely inspired by keeping up with the competition, but really that is not an exclusively unique feature of the KPA either. And you folks glossed over the other point I made that Fractal now allowsyou to choose the FW you want to run for amp modeling so you can update the unit to get other features but choose whether or not to take advantage of modeling enhancement.
To anyone new to this board, IMHO there are a handful of folks over here (not pointing at you viabcroce) with a bit of a grudge towards Fractal and TGP (more specifically to 1 or 2 members there) who use this board to voice their displeasure.
It makes sense that you try to make your product as good as possible even if you have to take ideas from some other products. I think the issue for some people is that for them it does not make sense to publicly bash the main functionality of product, calling it inferior and then try to copy it, or take as many ideas from it as you can and use them to improve yours.
Saying that the tone Matching feature on the surface does seem inspired by the Kemper is falling a bit short. Maybe you are not aware of it, but as soon as the KPA was released they announced their "own profiling feature". Tone Matching was the result of it, and they did not go further because they could not since Kemper has a patent on it. After that there has been several others "upgrades" influenced by the KPA. The latest is obviously another one. It seems to work by basically using stats (achieved by some sort of profiling) to correct modeling on the fly.