The principle of uv as much as I understand is to not tinker with virtual environments. With commands like uv run, uv sync etc you shouldn't have to worry about the virtual env and whether it's activated etc. if you are in a project, it will automatically pick up the right Python version . However if you want to do something not possible through those commands you can still get down to the virtual env. Plus for auto complete in IDEs you generally need to select the venv

I use this vscode extension called "python envy" that changes venv automatically depending on which folder of a monorepo you are in, it's great! Wish vscode had it built-in

Never heard of it, will try. Thanks

You install uv not python, you use "uv run <file>" instead of "python <file>", you dont think about anything else.

There's only one kind of Python virtual environment, and both uv and the standard library `venv` module make them (as does third-party support like `virtualenv`). The differences are in what gets put in them (aside from the actual packages you explicitly install), and in the interface for configuring that. In particular, the standard library defaults to bootstrapping pip, but you can easily skip that. And of course uv does not bootstrap pip, because it does pip's job (and much more).

If you are working in Data Science/ML its the best bet for handling dependencies in your project compared to the rest of the tools. Its use has exploded, especially because you can do ‘uv pip whatever’ if you insist on using pip.