I have had a love/hate/love relationship with Tailwind.

When Tailwind was first released, it generated a lot of buzz, and I naturally gave it a look. It was an intriguing notion—to define a multitude of tiny CSS utility classes that you embed directly in your HTML, giving you fine control over every tag. It was super cool.

However, I’m a huge believer in the separation of concerns. You shouldn’t mix your chocolate and your peanut butter, but it soon became apparent that Tailwind was asking me to do exactly that. One of the main purposes of CSS was to allow you to separate out the HTML and the code that styles that HTML. Didn’t Tailwind do the opposite? You can’t separate your concerns and have your design elements embedded in your HTML, can you? Well, no.

But the nature of web design has changed since CSS was first built. Most frameworks, whether it be Angular, React, or Astro, have become component-based. But even those components were designed to separate CSS and HTML. For instance, in Angular, a component consists of three files: a TypeScript file, an HTML file, and a CSS file.

But those components are becoming more and more granular. At the same time, the look and feel of websites have become more standardized. Button colors, for example, have standardized so that blue means “you can trust this button” and red means “be careful when pushing this one.” So the need for customized colors has been reduced.

Now here is where Tailwind shines. If you want standardized colors, Tailwind can define them. And if your colors and shapes are standardized, then Tailwind’s small utility classes that define those styles are useful. Finally, if those components are compact and self-contained, do you really need to separate your HTML and your CSS?

Ultimately, Tailwind is powerful and easy to use. Thus it has become very popular, if not a standard way to style websites.

And now Tailwind’s popularity might be its downfall.

This past week, the Tailwind team laid off 75% of their developers. Why? Well, according to Adam Wathan, the creator of Tailwind and the founder of Tailwind Labs, the layoffs were necessary because AI has caused the company’s marketing pipeline to dry up. Tailwind has that wonderful feature—the MIT License—which makes it basically free to use. Tailwind Labs depended on traffic to their website to drive “lifetime license” sales and sponsorships. But since AI now is doing more and more coding, developers don’t go to the Tailwind site and thus don’t purchase or support as much as they used to.

I kind of hate that.

Don’t get me wrong. I’ve written enough about agentic coding over the last few months to strongly support my bona fides as a vibe coder, but this is a real, live example of what can—and will—happen. We’ve seen Stack Overflow questions dwindle to practically nothing. Now AI is making it hard for Tailwind Labs to make money.

That’s the part I hate. Is AI simply going to make writing new code and frameworks not worth the effort? If so, then where will new code and frameworks come from?

I suppose that the answer to that is agentic AI itself. But only time will tell if AI can take over the task of creating better frameworks and libraries for our (its?) use, or if we will need to come up with a newer, better model for making human-generated libraries profitable.

I love Tailwind and I love agentic AI, but I hate what is happening to the former because of the latter. Who is going to build the future?