HTML-First Web Development Is Having a Real Moment
A post claiming that switching to an HTML-first site doubled users overnight is getting a lot of attention, though the thread quickly clarified that 'doubled users' really meant 'halved form abandonment,' which is actually the more impressive metric. The debate in the comments is sharp: some people credit Remix for reviving the pattern, others push back saying Remix isn't popular enough to claim that, and a fair number of skeptics say that people who built bad sites in React will build bad sites in Astro too.
A separate thread on CSS 'unavoidable bad parts' is circling the same territory. A commenter with the best ratio of dunks to insights points out that the author admits to never having written production CSS before declaring you shouldn't use CSS selectors. Both threads are really about the same thing: frontend complexity has ballooned past what most teams can manage, and there's a genuine backlash forming.
The pattern: performance and simplicity are winning arguments again. The 'just use less JavaScript' camp is gaining converts not because of ideology but because the data, reduced abandonment, faster load times, is starting to pile up. This is a slow-moving trend that's accelerating.
So what?
If your web product has high drop-off at sign-up or key conversion points, auditing your JavaScript bundle size and switching to server-rendered HTML for critical flows is a concrete, low-risk experiment with a documented upside. You don't need to rewrite everything; start with the funnel.