[javajosh]

It's important not to ignore the possibility that perhaps front-end dependencies are out-of-hand, and need to be reduced. ESM cannot fix a decade of bad practices enabled by front-end build bundlers. ESM isn't there to be a viable alternative to webpack. It's there to enable a different vision of application deployment where apps are smaller, and javascript gets css's transitive import() sub-resource distribution, avoiding the headache of a linear list of global scripts.

I really like ESM because I like where it's trying to steer the community of browser application builders. I think front-end builds are terrible on many levels, not the least of which is the obfuscation of code that undermines one of the best features of the web's software distribution, which is its openness. And another major benefit of webapps is that none of the front-end languages require a build step! This makes iteration very fast; if you can make do without the the safety net of a compiler, you can enjoy the speed of not using the bundler.

[crooked-v]

I think the basic problem with that comparison is that people use bundlers with CSS all the the time for exactly the same basic-physics-of-round-trips reasons. It's common for `import()` in CSS to end up exactly what it usually in frontend JS, a dev organization tool rather than something actually exposed to the browser.