The Rise of Vibe Coding — A Designer's Take
Vibe Coding · AI Development · Designer's Perspective · 2026

The Rise of Vibe Coding — A Designer's Take

Author Manav Tyagi
Published Jun 02, 2026
Read Time 10 min read

A new term has entered the tech lexicon — "vibe coding." It's the practice of building software not by typing code, but by describing what you want to an AI and letting it write the code for you. As a designer who cares deeply about the craft of building digital products, I have a lot of thoughts about what this means — for creativity, for quality, and for our collective future.

The Phenomenon

What Is "Vibe Coding" — And Where Did It Come From?

Quick context: The term was popularized by Andrej Karpathy's February 2025 post. Since then, the concept has gone from niche developer experiment to mainstream product strategy. Startups are now building entire products with vibe coding workflows, and the conversation has moved well beyond "is this real?" to "how do we do this well?"

Why It Matters

Why Designers Can't Ignore This Shift

  • The Line Between Design and Code Is Dissolving
    Vibe coding doesn't eliminate the need for design thinking — it eliminates the technical barrier to expression. A designer who can articulate their vision precisely now has the same implementation velocity as a developer. That's a profound change in creative power.
  • Iteration Speed Has Become Superhuman
    Traditional design iteration: wireframe → feedback → high-fi mockup → dev handoff → review. Vibe coding iteration: prompt → working prototype → adjust prompt → done. The feedback loop that used to take days now takes minutes. This changes how we test ideas, how we collaborate with stakeholders, and how we validate assumptions.
  • Non-Designers Are Now Building Interfaces
    This is the part that stings a little. When a startup founder can vibe-code a product in a weekend, your value as a designer has to be more than "I know how to use Figma." It has to be about the thinking — the user research, the systems logic, the accessibility decisions — that AI still can't replicate.
  • Design Intent Is Becoming the Scarce Resource
    In a world where execution is democratized, the people who know what to build — and why — become exponentially more valuable. Clear design intent, articulated in precise language, is now the fuel that powers vibe coding workflows.
The Opportunity

Where Designers Have a Genuine Edge

Design Thinking & Problem Framing

  • Defining the right problem before building the solution
  • Translating business goals into user-centred outcomes
  • Questioning assumptions that non-designers never consider
  • Structuring research to surface non-obvious insights

Accessibility & Inclusion

  • Ensuring colour contrast, focus states, and ARIA roles are correct
  • Designing for cognitive load, not just visual appeal
  • Recognizing when AI-generated UI creates exclusion patterns
  • Advocating for users who don't fit the default assumption

Systems Thinking

  • Building coherent design systems — not a collection of one-off screens
  • Maintaining visual and interaction consistency across scale
  • Designing tokens, states, and edge cases that AI misses
  • Thinking in patterns rather than individual components

Taste, Critique & Direction

  • Knowing when a vibe-coded output is "good enough" vs truly good
  • Articulating design feedback that improves AI outputs
  • Understanding emotional resonance and brand personality
  • Seeing what's missing — the spaces between the pixels
The Honest Critique

What Vibe Coding Gets Wrong — And Why It Matters

  • The Accessibility Blindspot
    AI-generated code is often visually impressive but accessibility-poor. Missing ARIA labels, incorrect focus management, insufficient colour contrast — these issues are invisible to someone who isn't trained to spot them. And they disproportionately affect users with disabilities. Vibe coding without accessibility review is vibe coding that excludes real people.
  • Aesthetic Convergence — Everything Looks the Same
    When thousands of teams prompt AI with similar language, they get similar outputs. The dark sidebars, the card grids, the gradient buttons — these patterns are everywhere because AI is trained on patterns. Distinctive, brand-specific design requires human art direction that breaks from the mean. Otherwise, we're all shipping the same product in different colours.
  • Hidden Technical Debt
    Vibe-coded products can accumulate technical debt rapidly — inconsistent naming conventions, duplicated components, bloated bundles, poor performance on slow connections. This isn't immediately visible in a demo but compounds into real problems at scale. Design systems thinking is what prevents this — and most vibe coding workflows skip it entirely.
  • No User Research Baked In
    Vibe coding can ship an interface fast. It cannot tell you whether that interface solves the right problem for real users. The risk of building beautiful solutions to the wrong problems has never been higher. Speed of execution without research foundations is how you ship confidently in the wrong direction.

Worth sitting with: The most dangerous version of vibe coding isn't bad code — it's overconfident code. When the output looks polished, it's easy to assume it's right. But visual fidelity and design quality are not the same thing. As designers, we know this. Not everyone does.

The Practical Guide

How to Use Vibe Coding Well — A Designer's Framework

Use Vibe Coding For… But Apply Design Judgment On…
Rapid prototyping and concept exploration Validating that the concept solves a real user need
Generating component scaffolding quickly Ensuring accessibility, semantic HTML, and naming conventions
Iterating on layout variations in minutes Choosing which variation best serves the user's mental model
Translating design system tokens to code Maintaining consistency and catching token drift
Generating copy, icons, and placeholder content Reviewing for brand voice, inclusivity, and tone alignment
Shipping a first draft to get real feedback faster Treating feedback as design research, not just bug reports

The mindset shift: Stop thinking of yourself as "a designer who doesn't code." Start thinking of yourself as a design director who now has an extremely fast junior developer available 24/7. Your job is direction, taste, critique, and quality control — not manual execution.

The Ecosystem

The Vibe Coding Tools Worth Knowing in 2026

  • Cursor — The Designer-Developer Bridge
    Cursor is a code editor with an AI that understands your entire codebase. For designers who have some comfort with HTML/CSS, it's the closest thing to having a developer co-pilot inside your creative environment. The contextual awareness means you're not just generating isolated snippets — it understands your project's structure.
  • Lovable — Designs That Actually Work
    Lovable lets you describe apps in plain language and get working React applications back. What makes it notable from a design perspective: the outputs are actually polished. Not pixel-perfect, but usable. Great for validating concepts with real stakeholders rather than static mockups.
  • v0 by Vercel — UI Generation with Design Sense
    v0 is specifically oriented around generating component-level UI. It outputs clean, Tailwind-based React components that are easy to integrate. For designers working with developers, it's a wonderful way to communicate intent — generate the component, hand it over, and let the developer adapt it to the production codebase.
  • Claude Design & Google Stitch — The Creative Suites
    These are the tools most explicitly built for designers. Claude Design ingests your brand system and generates within it. Stitch understands multi-screen flows and visual context. Both represent a new generation of AI tools that speak the language of design — not just code. Watch this space closely.

My Take: The Vibe is Real — But So Is the Craft

I've been a designer long enough to remember when Sketch replaced Photoshop, when Figma replaced Sketch, when no-code tools promised to replace everything. Each shift changed the tools, but it never replaced the thinking. Vibe coding is the next version of that story.

What's genuinely different this time is the pace. The gap between idea and artifact has never been shorter. When I can describe an interface and have something to click through in five minutes, the entire design conversation changes. I'm not explaining what might exist — I'm showing what does exist, right now, and refining from there.

That changes how I collaborate. It changes how I present ideas. It changes how I validate assumptions before investing hours in pixel-perfect mockups. And honestly? It changes how I think. The constraint of "I have to draw this" becomes "I have to describe this clearly" — and clear description requires clear thinking.

But I'll be honest about what worries me too. The design community is at risk of confusing speed with quality. Vibe coding can ship fast. It cannot ship thoughtfully without a designer in the loop who asks: Does this actually serve real people? That question — and the discipline to sit with it — is still entirely human.

My view: lean into vibe coding as a tool for exploration and speed. But don't let the vibe replace the craft. The best design work I've seen from AI-assisted workflows isn't the code the AI wrote — it's the clarity of thinking that went into the prompt. That clarity comes from doing the design work properly. It always has.

The vibe is real. The craft matters more than ever.