I think the truth is to be found somewhere in between.
This debate is really about what people want and how they envision the Web. Either as hypertext documents that link information; or as a platform for "rich web applications."
While people argue over the problems JavaScript pose when building applications. The same type of arguments also pop up when it comes to CSS. How the focus on semantics hampers building rich applications with a dynamic DOM structure.
I posted a comment a few days ago that goes into detail about this. I hope you will forgive me for repeating it here. I feel it's relevant to this discussion as well.
What changed over the past 30 years is how the Web evolved from a small group of academics and engineers 30 years ago into the billions that it connects today. It's not just hypertext. It's hypertext and every other conceivable business case which is being shoehorned into the browser.
The Web originally wasn't conceived to provide affordances that allow for "rich media experiences". Foundational technologies - HTTP, HTML, CSS, JS,... - weren't originally designed to build complex user interfaces and applications. Yet, as more and more and more people got connected: that's exactly what they wanted.
Remember Flash? Or Java applets? Or Silverlight? Those were hailed as additional technologies that would merge complex user interfaces into the Web. Macromedia coined the term "Rich Web Applications" back in 2002. All of them failed for various reasons.
Ever since the Web Standards moving of the mid-00's, communities of designers have taken CSS and HTML and tried to bend and use them in any way possible to bring visually stunning experiences on the Web. This is where you will find the direct roots of many ideas and dynamics that are currently in vogue. It's their influence, as well as other dynamics - e.g. the introduction of the smartphone, the rise of big tech - that has defined how browser technology has evolved.
Over the past decade, browser vendors have tacked on additional browser API's on top of the original, decade old foundations, while communities of frontend developers have spawned ever intricate constellations of interlocking JavaScript libraries and frameworks that specifically try to break out of how the Web originally was intended to work.
Some things have become easier as specifications evolved a bit (e.g. CSS3). But at the end of the day, your browser still needs to download HTML, JS and CSS, parse all of that, execute and render it onto a canvas.
Tailwind solves a particular issue: it tries to do away with the semantics of CSS and abstracts how presentation relates to the DOM structure in order to allow developers to focus on what they should be building: rich, highly interactive applications.
Tailwind, however, isn't the right tool if you want to focus on building classic hypertext documents / websites. It's still totally valid to build things using the original web standards principles.
It's not a tale of one or the other. It's a tale of both are valid business cases. I think that should be the core take away here.