JavaScript seems to be unique in that you want your code to work in browsers of the past and future—so a lot of bloat could come from compatibility, as mentioned in the article—and it's a language for UIs, so a lot of bloat in apps and frameworks could come from support for accessibility, internationalization, mobile, etc.
Lots of developers don't even say they are JS devs but React devs or something. This is normal given that the bandwidth and power of targets are so large nowadays. Software is like a gas, it will fill all the space you can give it since there is no reason to optimize anything if it runs ok.
I've spent countless hours optimising javascript and css to work across devices that were slow and outdated but still relevant (IE7, 8 and 9 were rough years). Cleverness breads in restrictive environments where you want to get the most out of it. Modern computers are so large that its hard for you to hit the walls when doing normal work.
Every C++ app I install in linux requires 250 packages
Every python app I install and then pip install requirements uses 150 packages.
Not to the language but its users. Not to bash them, but most of them did not study IT on a university, did not learn about the KISS principle etc.
They just followed some tutorials to hack together stuff, now automated via LLM's.
So in a way the cause is the language as it is so easy to use. And the ecosystem grew organically from users like this - and yes, the ecosystem is full of bloat.
(I think claude nowdays is a bit smarter, but when building standalone html files without agents, I remember having to always tell chatgpt to explicitely NOT pull in yet another libary, but use plain vanilla js for a standard task, which usually works better and cleaner with the same lines of code or maybe 2 or 3 more for most cases. The standard was to use libaries for every new functionality )
It sure seems like it is because JS devs, by and large, suck at programming. C has a pretty sparse standard library, but you don't see C programmers creating shared libraries to determine if a number is odd, or to add whitespace to a string.
Believe me, if C had a way to seamlessly share libraries across architectures, OSes, and compiler versions, something similar would have happened.
Instead you get a situation where every reasonably big modern C project starts by implementing their own version of string libraries, dynamic arrays, maps (aka dictionaries), etc. Not much different really.