Yes, I was a mentor for Stanford Machine Learning for a few years as well, so I am familiar with how things worked over there. But remember all the hassles with how to get Octave or MATLAB installed and properly configured on your system. Those are non-trivial issues and just provide more “friction” in the process. That’s why the Jupyter notebooks are a step forward: the entire course environment is web based and does not depend on the user’s environment other than having a reasonable browser. You can even use the course from a tablet, not a real computer. If you want to run things locally, you can, but then you will deal with all the issues of getting things set up on your local system in a way that is compatible. I completely understand why the course staff does not want to get involved in that. My prediction is that they will politely decline your request to go down that path.
More recent update: they have provided an easy way to download all the files in a given assignment. Please see this thread. But note that getting the assignments to run locally is not a simple matter and requires a lot of software to be installed. Here’s a thread about how to get started down that road. And here is a thread about how to run the notebook on Google Colab.