I wonder if there's a worthwhile time-saving in aggregate? To me it seems like overall the bottleneck is always thinking what code to write/edit, not the actual edition. So I'm not convinced shaving a couple seconds here and there outweighs the benefit of a modern IDE or offsets the time spent ricing such a setup.
It's not mainly about speed (although the ceiling is much higher), but about ergonomics, and, plainly, enjoyability. The other benefit of making your own environment is mastery and control; only you can decide when and how the environment changes.
> To me it seems like overall the bottleneck is always thinking what code to write/edit
I feel the same way. But small advantages compound to at least some extent (see e.g. https://danluu.com/productivity-velocity/); and I find that noticing that it takes more time than I'd like to write/edit something, risks breaking my flow.
Seems like the opposite to me. IDE fashion changes like the wind. "Normal developers" are constantly starting over with new tools and new paradigms every time they start work on a new project. Gotta have the right vscode extensions to flash your firmware, IT only supports IntelliJ on the new job, ad infinitum. It's been this way for decades.
My .emacs file gets updated regularly, sure, but it's thirty years old, and my basic flow hasn't changed.