I used to work on an single page application that was pretty complicated with heavy DOM updates. In the team, we wanted to investigate at the time if React could help us with some performance bottlenecks. However, I remember very much with that buy-in with JSX didn't appeal much to me. As we started to write more code against it, it was okay.

On reflection though, I don't think React would have helped to solve the problems we thought we had. I only tried to see it as a solution to a problem, not seeking to use it to be part of a sub-industry around this library. But, even as I write this, I get a recruiter email asking for React experience. I have not followed it much for a few years, but using React now doesn't seem to fit the exact problem it tried to originally solve anymore (if that problem even matters now) but is a completely separate thing. But I really don't care anymore, because I don't work in any focused frontend web development now.

I wrote something in plain JS a couple of weeks ago, up to a point where it was generally useful for myself for my work. It was wonderful - no crazy complex toolchain, no React to deal with, just html and some JS to do stuff. Doing the simplest thing in a new project (no extraneous toolchains if you can help it specifically) should be the first thing to do. This is not a fad, it's just taking stock in the time you have to make usable programs.