A designer hands over a frame. A developer asks what happens when the menu opens. The answer is somewhere between “good question” and another meeting.

Designer over here.

Developer over there.

Handoff in the middle.

That model becomes less convincing when design agents can work directly on a canvas, coding agents can read design systems, running interfaces can be captured back into Figma, and prototypes can become functional long before a formal handoff begins.

I think this creates an advantage for people who understand both sides. That is a professional judgment, not a measured pay premium for hybrid roles.

A designer who understands code can judge whether an interaction is structurally sensible before it reaches development.

A developer who understands design can recognise when technically correct output is experientially poor.

A hybrid practitioner can use AI across the entire loop.

That does not mean everyone should become equally expert in everything.

Breadth without depth is still a problem.

But the ability to cross boundaries becomes more valuable when AI handles more of the mechanical translation between disciplines.

Figma’s current direction makes this explicit. MCP connects structured design context with coding agents. Code-to-canvas workflows bring production UI back into design. The design agent can act directly on the canvas.

In that environment, the key skill is increasingly direction.

Can you tell whether the generated component respects the system? Can you understand the layout and the DOM? Can you identify why the mobile experience fails? Can you adjust the design rather than merely accepting the first generated option?

This mirrors the kind of work that often happens in real projects.

A responsive GitHub contribution calendar is not purely a design problem or purely a coding problem. Neither is an iframe that breaks at mobile width, a navigation system with awkward state, or a CMS component whose content model makes the interface harder to maintain.

The boundary is artificial.

AI is making that more obvious.

A more resilient career strategy may be to build judgment across the product, while keeping real depth in a craft.

It is to become the person who can understand the whole problem well enough to direct the model.

06GitHub context

HZ-Archives is an example from my own design and development practice; verify the individual commits before citing them as public evidence: https://github.com/Hadizainal/HZ-Archives/commits