I have read the paper.
I do not find it surprising that different opamps sound different.
The question is which opamps were tested. Were they old, were they new, were they cheap, were they expensive? Which ones measured best? Which ones measured worst? Which ones were preferred? What was the correlation, if any, between age, cost, measurements and other factors, and preference?
I do think it is interesting that the study concluded that opamps which had more distortion were preferred. What was the nature of the distortion in those opamps which were preferred?
The paper would appear to suggest that chasing the best possible numbers does not always result in the "best" possible sound, where "best" is determined by preference.
Perhaps all of those Harman DBTs have resulted in amplifiers and speakers being unleashed on the market which were not those that had the best measurements, but which nevertheless were preferred by trained listeners.
Some may be interested in owning equipment which has the best possible measurements regardless of how it sounds. Others may be interested in owning equipment which sounds best to them, regardless of measurements. Think tube vs. solid state, analog vs. digital, USB vs. S/PDIF, negative feedback vs. no negative feedback.
Do you want to own a DAC or use a digital interface which has lower jitter but sounds worse than a DAC or digital interface with higher jitter?
Should you offer equipment for sale to your customers which has better measurements or which sounds better? Should they be offered a choice?
Which equipment has better performance - equipment which measures better or equipment which sounds better?