There are basically three camps, as I understand it, who all agree this is impossible.
1. Gil Kalai and friends who buy QM but says there is "correlated-noise" that will limit scale even with perfect...
I tend to agree with your conclusion. Talking about and generating new ideas is fine, but until seriously threatened, making any changes now or bip’s without thoughtful application of NIST-tested methods, etc., it’s primarily an academic exercise and shouldn’t be anywhere near bitcoin code.