ETTR: Is it still a thing for you?

Depends how one got there and where one is looking:

1) If the desired aperture and shutter speed would have resulted in no desirable data being recorded in the raw data above 1/4 of full scale at base ISO and the photographer had the artistic latitude to change aperture and/or shutter speed by 2 stops, then the image's SNR would be less noisy, for instance by a factor of two in the brightest highlights.

2) If on the other hand aperture and shutter speed were already maxed out for artistic reasons and the ISO had been increased by 2 stops to maximize the signal, then it would depend on the camera. Canon users would be much, much better off, while Nikon users (especially those with Exmor sensors) only possibly a little.

Are you sure? Let's look at an example, the situation in 1) above. Brightest raw highlights according to artistic intent were going to be recorded around value 4000, but the photog decided to increase shutter speed and/or aperture by 2 stops to maximize the signal, recording them instead around raw value 16000 and then bringing them back in post.

Signal = 4000, shot noise SNR = 63

Signal = 16000, shot noise SNR = 126

after -2EC correction in well behaved raw converter:

Signal = 4000, shot noise SNR = 126

Hmm, by maximizing the signal the photog got 1 stop better noise performance in the highlights, well done.

Perhaps, but only in the darkest shadows.  Noise is not a separate entity - it is a standard deviation superimposed on the mean signal.