Howard makes the case for why even "real" programmers should give notebook environments a chance. In addition to supporting literate programming ("code as literature") and exploratory programming ("code as scientific notebook") in a live coding environment, Howard explains how notebooks can improve documentation, learning and sharing, testing, and deployment. And add-ins and tools, like Howard's own nbdev, can help address what's missing in Jupyter Notebook. As an example, Howard notes how fastdoc even enabled him to write and publish "Deep Learning for Coders with Fastai and PyTorch: AI Applications Without a PhD using Jupyter Notebooks." Excellent, inspiring talk!
Jeremy Howard - Creating delightful libraries and books with nbdev and fastdoc | JupyterCon 2020 https://www.youtube.com/watch?v=HKt19-GsA1I