It's not a question whether it works. It's a question of whether it's being used. And it's not, really. Those code paths are less tested and less performant, so

It's not a question whether it works. It's a question of whether it's being used. And it's not, really. Those code paths are less tested and less performant, so why would you choose them even if you start a new project without baggage? Python 3 won't see any traction unless it offers a clear improvement over 2. Some people think that's async. I have my doubts.