It's largely because Python had excellent numeric libraries, like NumPy and SciPy. You could script Fortran linear-algebra kernels to crunch your numbers fast i

It's largely because Python had excellent numeric libraries, like NumPy and SciPy. You could script Fortran linear-algebra kernels to crunch your numbers fast in a high-level language, and that's a killer app in the ML space.