The Great Design Reset

Stop Mourning Your Figma Files

The era of the "perfect screen" is over.

Today, product design is a bloated bureaucracy of artifacts. We became pixel lawyers, negotiating corner radii and padding in endless review cycles. We convinced ourselves that "craft" meant maintaining a 2,000-component Figma library that no one else cared about.

Now, AI is here, and the collective panic is justified. "If an LLM can generate the UI, what do I do?" "If users just talk to the computer, do interfaces disappear?"

The freak-out is valid, but it's misplaced. You're mourning the wrong thing. You're mourning the loss of the deliverable, not the loss of the job.

If your identity was wrapped up in drawing rectangles, yes, you're in trouble. But if your identity is about deciding how a system behaves, how it handles failure, and how it serves human intent, you've never been more valuable.

The "freak-out" is happening because we're in the middle of a phase shift. We're moving from Deterministic Design (where you draw every state) to Probabilistic Design (where you define the rules of behavior).

The pixel factory is closing. The behavior factory is opening.

Beyond the Chatbot: Liquid Interfaces

The laziest take on AI is: "Everything will become a chatbot."

Bullshit. Chat is just terminal or command line for regular people. It's high-friction, low-discoverability, and cognitively exhausting. Nobody wants to talk their way through a complex dashboard or a video editing timeline.

The future isn't chat. The future is Liquid.

We're moving toward interfaces that assemble themselves for a specific moment and then adapt or vanish.

Imagine a travel app. Today: You navigate a maze. Search → List → Details → Checkout. I designed those screens six months ago. They're frozen.

Tomorrow (Liquid): You say, "Book a trip to Tokyo for me and my partner, focusing on food." The system doesn't just show a list. It spins up a custom mini-app on the fly. It generates a comparison table because it knows you're price-sensitive. It pulls up a map overlay of restaurants because you mentioned food.

This UI didn't exist ten seconds ago. It won't exist ten seconds from now. It was generated for your specific intent.

This is UI on the fly. And it changes everything.

• We stop designing screens. We start designing components that can assemble themselves.
• We stop mapping "happy paths." We start mapping intent boundaries.
• We stop polishing pixels. We start polishing logic.

The interface of tomorrow isn't a painting; it's a living organism. It breathes. It adapts. And someone has to teach it how to adapt. That's you.

The New Role: Architect of Behavior

So, what do you actually do all day?

If you can't spend 40 hours tweaking a modal, what fills the time?

You become a System Architect of Behavior.

In a world where AI agents go off and do things like "triage my email" or "negotiate this refund," the UX is no longer about the button color. The UX is about Transparency, Control, and Trust.

Your new deliverables are:

1. The Behavior Brief: A document that defines what the AI is allowed to do. Does it ask for permission before spending money? Does it sound confident or humble? What topics does it refuse to engage with?

2. The Contract: The "handshake" between human and machine. How does the user know the AI is working? How do they stop it? How do they undo a mess?

3. The Guardrails: The safety nets. The AI will lie. It will hallucinate. It will try to do things you didn't ask for. Guardrails are designed constraints that prevent the AI from doing the worst-case thing.

You are no longer the decorator. You are the zoning board. You are the safety inspector. You are the choreographer.

This is a promotion, not a demotion. You're moving closer to the core. You're moving closer to the product.

The Syllabus

How do you actually pivot?

1. Kill the Mockup as the Source of Truth. Stop treating Figma as the final product. Figma is a sketchpad. The code is the truth.

2. Learn the Material. You don't need to be an ML engineer. But you need to know the difference between a context window and a vector database. You need to know why RAG hallucinates less than raw GPT.

3. Design for "Maybe." Traditional design is binary: Success or Error. AI design is probabilistic: "I'm 70% sure this is right." Design the UI for that 70%.

4. Become a Specialist in Human Handoffs. The most critical moment in any AI system is when it fails. The "Human-in-the-Loop" pattern is your bread and butter.

5. Focus on Failure States First. Most designers do it backward. They design the success case first and treat errors as an afterthought. In AI, the error is the feature.

You have a choice

Path 1: Become a Decorator. You learn Figma better. You specialize in "aesthetic UI." In five years, you're out of work, because AI is better at beautifying things than you are.

Path 2: Become an Architect. You stop worrying about pixels. You learn systems thinking. You learn to code—or at least to prototype in real tools. You learn the constraints of AI. You learn to design behavior. In five years, you're running the product.

If you want to keep working, I’d suggest going with the first ;-)