On 2 totally agree!
On 3 how is log(exp(x)+1) an alternative to softmax? The outputs are not class probabilities. But I agree in general. I have many many many problems with the use of Softmax and hope there is a better alternative.

I will say I am a little bit skeptical of this field but here is an example: https://arxiv.org/pdf/2306.05415.pdf

And a survey:

https://arxiv.org/pdf/1911.10500.pdf