CSS: Unavoidable Bad Parts

Most of the useful pushback landed on one point: the article reads like advice for simple document pages, not for modern app UI, marketing sites, or component systems. People were blunt that wrappers, classes, and some presentation-driven HTML are often the practical cost of building complex layouts that survive real content, responsive states, and composition inside larger components. The consensus was not that div soup is good. It was that modern CSS still does not fully let you separate structure from presentation in the clean way people once promised, so markup ends up carrying some layout responsibility. A second theme was that the article undersold how much CSS has improved. Flexbox and CSS Grid were repeatedly cited as having removed a lot of the old pain, with newer tools like `gap`, `subgrid`, `display: contents`, `:has()`, variables, and nesting changing what is possible or maintainable. That did not make commenters claim CSS is elegant. It made them say the author's framing felt stale. Several also drew a line between “same HTML can adapt to many viewports,” which CSS mostly does well, and “HTML and CSS are fully orthogonal and swappable,” which people said never really scaled beyond simpler content sites. The practical advice that got traction was narrower and more useful than the post's broad verdicts. Use relative font sizing for accessibility, but do not blindly tie every padding and spacing value to `rem` or larger user text can make mobile layouts worse. Keep wrappers when they simplify composition, alignment, or accessible form controls, but remove gratuitous nesting that forces hacks like `height: 100%` through multiple layers. And be cautious with features that increase selector complexity or specificity, because many teams still find maintainability breaks there before raw layout capability does.