Design Systems Are Dead (And We Killed Them)

Yep, all those meticulously curated repositories of buttons, cards, and well-documented spacing tokens are dying.

By this point, many of us have spent years or months building the perfect and completely flawless design system. We've evangelized it, polished it down to every minute detail, written documentation that nobody read, and maybe plunged ourselves into an almost-existential crisis deciding between 8px vs 12px padding.

But it might be that all along, we have been optimizing for the wrong goal.

A Consistency Trap

Somewhere around 2015, the design community decided that "design systems" were the answer to every problem. Building out extensive-ass Figma libraries with 34+ button variants, documenting every spacing token, and preached “design consistency” like it was the word of God.

The promise was kinda seductive too: build it once, use it everywhere. Consistency at its best.

And it worked. Maybe a little too well.

Now every SaaS product looked identical. Stripe. Airtable. Notion. Linear. Same rounded corners. Same muted palettes. Same generous whitespace. Same ghost buttons.

We didn't build design systems. We just built design uniformity.

When was the last time you used a product and thought, "Wow, this interface has personality"?

We optimized and literally dragged the soul out of every design.

The Shift Nobody's Talking About

The new AI hype is making static component libraries feel...just useless?

Not because AI gives some magical optimizations to the design. But because it exposes a fundamental flaw in how we approached design.

Why were we maintaining 12 button variants when we could define straightforward rules for when buttons should adapt based on context, user state, device, and intent?

We've been creating artifacts (Figma files, Storybook docs) when we should've been creating intelligence.

The future isn't about maintaining components. It's about defining logic that generates the right UI at the right moment.

But here's the catch: AI can optimize, but it can't create.

What Dies, What Lives

What's becoming obsolete:

• Static, hand-maintained component libraries

• One-size-fits-all solutions

• Design systems as documentation projects

• Treating consistency as the highest virtue

What's still essential:

• Design principles (the why behind decisions)

• Accessibility requirements

• Brand essence (not guidelines – essence, the heart and soul of a brand)

• User empathy and understanding

• Creative expression

• Systematic thinking (not systematic artifacts)

The difference matters. Systematic thinking is timeless and permanent. But that 60-page Figma file? That's temporary.

Evolving Roles of a Designer

So if we're not going to maintain component libraries, what are we going to do?

Well let’s just say we're becoming architects of logic and emotion.

Instead of creating several unnecessary and totally irrelevant button variants, you define the rules that decide which button works best based on user context. You're designing systems of decision-making.

And here's where humans stay irreplaceable: injecting humanity.

AI can come up with a visually perfect interface. It can't generate the feeling behind it. It can't make someone care about a product or experience instead of just tolerating it.

Our job is to understand the user’s behaviour and their needs deeply enough to translate it into experience. That's the work AI can't even begin to touch.

The Uncomfortable Questions

If you're feeling defensive, ask yourself:

Are you maintaining a design system because it genuinely helps users? Or because it just sounds like "best practice"?

Is your component library enabling creativity or constraining it? When did someone on your team last build something new instead of assembling existing parts?

What would happen if you modified your design system tomorrow? Would your product improve or collapse?

Be honest.

Moving Forward

Design systems solved real problems – inconsistent experiences, poor design-engineering collaboration, and spaghetti codebases.

But like all solutions, these systems created new problems in their wake. Now the industry is just evolving in response to these newer problems.

The next few years might not be about maintaining libraries, but about understanding human behaviour deeply enough to create adaptive, emotionally relevant experiences.

Whether those are AI-generated, hand-crafted component-based, it doesn't matter. The thinking does.

What are you going to build next?

That's the tea, folks. Catch you later 👋