with conda in an built env I never had issues with the performance
(there could be windows more likely to be blamed 
But like the video demonstrated, in conda the sequence is key (and pinning packages and channels [before installation]-> makes it quite stable - because of that I used it in docker.
In my case the solver took hours and later in conda 4.x more than a day or failed (even for already built envs).
For me working with Jupyter without extensions would be like working with VScode without extensions…
I saw the documentation (which is quite old in some points) and tried some things out but for now I don’t see too much benefit from withJupyter (to get an advanced Data Science environment built)
… will try some more