Unfortunately it's misleading and misinterpreted.
The recently published underlying paper has been very popular with audiophools and other people who don't really understand it because things tend to be written about it in such a way that they seem to imply that humans beat math/physics. The truth is that what has been demonstrated is that humans do in fact beat a particular limitation involving a type of linear processing - but this was something that was basically known but apparently unconfirmed.
The basic idea is that there is a limit (known as the Gabor limit) in time vs. frequency resolution when performing a linear process (an FFT) and that it has now been demonstrated that humans can perform better than this limit.
But the idea that human brain does non-linear processing has been known/assumed for a loooooong time by people who work with human audio perception. Like say, for instance, people who do things like perceptual audio coding for things like mp3's.
Humans just don't use really a linear process like an FFT to perceive things. This isn't a new idea; it just (apparently) hasn't been scientifically confirmed before in this specific case.
But none of this matters to the listener, because though FFT's are used in audio processing, the limit doesn't adversely affect the audio itself - it's just puts a limit on how you can process it using linear math. You can either optimize for frequency resolution or time resolution, but not both at the same time.
What is happening here is that people who don't really have any background in or understanding of any of this stuff are starting from an interesting confirmation of what was long suspected about human audio perception being non-linear and jumping from that to ridiculous conclusions about non-existent impacts on real world audio.
Carry on.