I see Go lacks vectorized notation, like python’s “a[:, 0] = b[0, :]”.

I personally can’t see myself going back to non-vectorized languages when dealing with math heavy topics like neural nets. In Python, this is also heavily optimized by NumPy, so the code is many times vectorized at the SIMD level.

I didn’t find you explaining why use Go for this, so I assume you just did it for fun (why not?). However, I would like to know your thoughts on this and what could be done to mitigate / optimize these issues in Go.