Nate Meyvis

There are a few reasons for this, but primarily, modern IDEs are just so wonderful and, although each of Vim and Emacs can be customized forever, I preferred just taking what VS Code gave me out of the box (plus some tweaks). [2/N]

Moreover, I was doing a lot of machine learning and data stuff, and I was spending so much of my programming time either (i) in Jupyter Notebook / Lab or (ii) sitting and thinking that it didn't make sense to optimize for minimizing keystrokes Vim-style. [3/N]

But! Now I'm a different / more "traditional" kind of SWE and I'm really feeling the burden of those extra keystrokes. Getting into Vim mode in an IDE is very helpful. [4/N]

A few lessons: I think @wycats was exactly right in a 2011 interview when he said that (paraphrasing) *not* diving in headfirst is an underrated way to learn Vim. You can just ease into it, learning one thing after another, without paying huge upfront productivity costs. [5/N]

Relatedly, Vim keybindings in a more recent IDE is, I suspect, a good happy medium for a lot of people. And, as with the point above, this is something that you might feel pressure not to do if you're learning Vim and (thus) reading lots of documentation from Vim purists. [6/N]

Also, this isn't at all original to me, but it's quite striking: thinking in terms of a distinction between command and insert mode has changed, for the better, the way I think about writing (and not just writing code). [7/N]

Alternatively phrased: If you're a writer, it's nice to make more explicit to yourself how much metaphysics is encoded in traditional methods of word processing. [8/N]

Finally, the case of writing software (IDEs and so on) is a neat case study in technological progress. I'd love to read more about it if anyone has suggestions. [9/9]