I want to break this list of things down a bit;
CSS Grid / Flexbox - Flexbox was defined in 2009, and was added to browsers about 10 years ago and iterated since. It was literally in the first public version of Chrome. This is not some new tech that's blindsided everyone. Grid is a couple of years old.
React - First released 7 years ago and iterated since. Reasonably mainstream for 5 years. That's not a very dramatic change.
WASM - Around for a few years now in browsers but tooling for different languages is still patchy. Entirely unnecessary for 99.9% of web development. Most people are still in the "use it because it's cool" phase rather than "use it because it's actually useful".
extensions to the web API - Various, obviously. Can't really argue this point.
lots of new JS features - Various, but not actually very fast moving with new features. Most things are made available earlier through transpilers, so you can use a feature without it being in mainstream JS engines if you want to. Can't really blame the web if you make that choice though.
countless frameworks - You don't need to learn them all. Pick any number to use between zero and all of them depending on how much time you want to spend learning frameworks. Again, this is a choice, and if you're finding it overwhelming it's because you picked the wrong number of frameworks to learn.
plugins and build tools - Same answer again. Pick something and learn it. Update your skills every few years if you think your tooling is holding you back.
more homogeneity between browsers - This reduces complexity. To complain about that in a list of things to taking up your time is weird.
The web isn't that fast moving. I completely agree with your point about focusing on things that benefit the user though. I would argue that frameworks and build tools that enable me to iterate features faster do that though.