Modern GUI emacs has basically all of the features that the OP’s rant mentions (You can even run WebKit inside of emacs, if you want), but they have to be configured. It seems likely that they haven’t tried emacs in a while, because the rant demands things about packages that already exist. If this were a serious request for help rather than a rant, “Use Spacemacs or Doom-emacs” would be a fine response.
That said, there’s nothing wrong with a little rant now and then. There’s a pretty pervasive feeling among programmers that our tools for interacting with code certainly could be better, and probably should be. I assume the OP is hitting one of those moments right now.
More seriously, though, I believe that there is an answer to the headline question: we don’t have better code editors because the people with the combination of skill, desire, and opportunity almost always fall into one of the two traps:
1.) If they want to make a better general editor for experienced, skilled programmers, then they get sucked into learning an existing editor (usually emacs or vi, sometimes both) on the way, and they find that once you are proficient, those are really quite powerful coding environments.
2.) If you want to make a better general editor for people learning to code, then they tend to create something that is limited (simplified, streamlined, etc). Along the way, they either stop at a version that’s good enough (there are lots of these around), but definitely missing things that are important to (a smaller group of) skilled, experienced programmers – at which point, see #1. The strongest candidates survive, and some of them go quite far (Linus Torvalds uses a fork of micro-emacs, there are people at Google who use nano, etc.)
In the end, the “editor wars” between emacs and vi came down (IMNSHO) to a question between “fast and lean” vs. “expandable”. In the end, both became ”expandable enough”, and software growth (bloat) grew so extensive that both became “fast and lean enough”.