Use conda to set up new environments and switch between them, and use pip to install modules. It's really not as complicated as people make it seem. I haven't had any issues with this set up across Linux and MacOS, doing web / ML / data development for the last 3 years. It also avoids the pitfalls mentioned in the blog post.
I'll provide a witty reply to this comment as soon as conda finishes resolving the dependencies for my projects. Any time now...
Have you ever tried mamba[0]? It's much, much faster than conda for me.