I've been using this mapping for so long and it's so ingrained in my brain that I completely forgot about it. It used to happened that I'd start typing jjjjj into a textarea or a word processor, and it took me a while to realize what was wrong.

Doom Emacs had this behavior by default, and it caught me by surprise when it prevented me from typing "noto-fonts-cjk." I knew a lot of people do this in Vim, but it took some time to figure out exactly what had happened. Excellent framework though.

I tried this, but it (understandably) caused a delay whenever I typed j alone in insert mode

You can map both jk and kj to Esc, which will allow you to press both of them simultaneously. Then you can set your timeout quite low when in insert mode.

It only matters if you are tying "jj". If you are typing "just", it's just as fast and the "delay" is just visual.

Yeah, you get used to that.