I’ve found varied success. In general, I’ve encouraged the move across to being teaching source control. That has been in contexts where notebooks are being use

I’ve found varied success. In general, I’ve encouraged the move across to being teaching source control. That has been in contexts where notebooks are being used for critical outputs rather than exploration. When you get into MLOps as well, having .py templates actually makes the Data Scientist’s job easier as they can plug and play their models into a system that tracks inputs, outputs and changes for them