JS fatiques starts where web pages (not web apps) start to depend on thousands of kilobytes of javascript.
Now the default thinking is along the lines "I am starting a new project, I will use angular", instead of some analysis what kind of project this is, how much JS is needed and what is appropriate.
I've started chiding those at work who preemptively implement features before they're required, and not scoped. Commit the simplest code that will do the job, and if you like, create a ticket for a future sprint with cool new features.
At home, I wait until lack of a feature (like hot reloading) frustrates me, then block out a couple of hours to do it specifically. Otherwise you've got a flashy stack and no product.
