Yes, I understand the point that there is a limit in terms of SNR that current technology cannot help with. However, there are improvements with every iteration of a HA. I think these improvements help - and they matter. 25 years ago I could not go into a bar (with analog instruments). If music was on louder than 70DB everything was drowned out. 15 years ago, it was possible, but a struggle and I could not make out speech with clarity. 5 years ago it was a lot better, speech was delivered with clarity but I still had problems in noise. Now, it is a lot lot better. Yes, if it is 80DB and I am in a group - it is not a solution, but I can at least talk to people side by side or in front of me - even if I struggle with the group. As Jordan says, the HA cannot zone in on the talker of interest.
I am excited but I am also sceptical about hearing tracker ratings - because I don’t yet understand how they model the real world scenarios and how they know how to interpret the efficacy of each aid in terms of what comprehension will be delivered. It’s one thing to remove noise, but if the signal of interest is also removed, then comprehension doesn’t result. Therefore for me, anecdotal reports, subjective as they are, are very important.
If deep learning restores intelligibility then those who came up with it should be given the Nobel prize for science. That would be amazing.!