Is M1 support something people are working on? What are the issues / difficulties / magnitude of that project?
It seems everyone wants deep learning to work on M1 chips. At this point, I am quite sure there is enough interest for major frameworks to consider supporting it.
You'd think Apple would be doing their best to help with this, considering they market their "neural engine" pretty heavily.
Small addtion: matrix multiplications (and other operations implemented through BLAS) do use the M1's AMX matrix co-processor (through the Apple Accelerate framework).