Nice to see another player in the crowded hosted Jupyter notebooks space!
Still very far from smooth (<10) click experience.
The workflow took a while to get started:
Manually create account (no external auth option such as Github?) -> verify account -> create notebook -> set pw for notebook -> run notebook
I did all that only to find that no required libraries are there.... not even Pandas?
By comparision Google Colab onboarding is much smoother and you know you are getting a working Tensorflow, Pandas, Plotly etc out the box. Sure you pay the privacy tax to Google for this smoothness.
My biases:
- I've installed and maintain Littlest Jupyter Hub instance for teammates at my organization.
- I teach using local Anaconda installation.
- I use myBinder and Google Colab.
If I did not have a smooth experience I would imagine someone starting out would have a harder time.At the very least you should curate the packages so they run out of the box, without fiddling with requirements.txt, yaml, pip install, conda etc.