Brent Jackson says CSS utility libraries failed somewhat:
Eventually, you’ll need to add one-off styles that just aren’t covered by the library you’re using, and there isn’t always a clear way to extend what you’re working with. Without a clear way to handle things like this, developers tend to add inconsistent hacks and append-only styles.
I have a feeling Tailwind people would disagree. I have no particular opinion here, I’m just noting that Tailwind seems to have a more fervent fanbase than those early days of Basscss/Tachyons.
Brent goes on to say that CSS-in-JS solves the same problem, but in a better way:
CSS-in-JS libraries help solve a lot of the same issues Utility-based CSS methodologies were focused on (and more) in a much better way. They connect styles directly to elements without needing to name things or create abstractions in class selectors. They avoid append-only stylesheets with encapsulation and hashed classnames. These libraries work with existing build tools, allowing for code splitting, lazy loading, and dead code elimination with virtually zero effort, and they don’t require additional tools like Sass or PostCSS. Many libraries also include CSS performance optimizations, such as critical CSS, enabled by default so that developers don’t need additional tooling or even need to think about them.
No wonder people have been raving about this.
The one-step-back refers to the fact that CSS-in-JS is more open-ended and doesn’t encourage consistency as much. I’m not sure about that. Seems like if you’re building in a component-based way already, consistency kind of comes along for the ride, even before using design tokens and the like — which a CSS-in-JS approach also encourages.