Comment on "Harvesting information to control non-equilibrium states of active matter"

Abstract: In the article Phys. Rev. E 106, 054617 "Harvesting information to control non-equilibrium states of active matter", the authors study the transition from one non-equilibrium steady-state (NESS) to another NESS by changing the correlated noise that is driving a Brownian particle held in an optical trap. They find that the amount of heat that is released during the transition is directly proportional to the difference of spectral entropy between the two colored noises, in a fashion that is reminiscent of Landauer's principle. In this comment I argue that the found relation between the released heat and the spectral entropy does not hold in general, and that one can provide examples of noises where it clearly fails. I also show that, even in the case considered by the authors, the relation cannot be rigorously true and is only approximately verified experimentally.