How can it generate 3D worlds, videos, images and entire web pages, but still suck at front-end?

Nerdy Dev is a publication offering insights, tutorials, and resources for developers and technology enthusiasts. Readers can learn about programming languages, web development frameworks, and software engineering best practices. With articles, tutorials, and code samples, Nerdy Dev provides  guidance and expertise for building software applications and honing technical skills.

Adam Argyle

A critical take on why AI coding tools consistently underperform on front-end development tasks. The author identifies four core reasons: AI trained on outdated, template-heavy data; LLMs cannot render or visually perceive output; they lack understanding of architectural intent (SDD, BDD, state machines); and they have zero control over the chaotic browser environment with its endless permutations of viewport sizes, input types, user preferences, and browser versions. While AI handles boilerplate scaffolding and token migration well, it fails at bespoke interactions, pixel-perfect layouts, accessibility, performance optimization, and complex component states. The unpredictability of human behavior compounds the problem further.

Why AI Sucks At Front End · April 12, 2026

Because LLMs don’t have eyes or hands. They should fail at frontend, and they predictably do! Frontend is as infinitely complex complex as backend work, except perhaps more so because its so cross-discipline and multi-sensory.

Beauty is only skin-deep. And AI(s) don’t even have that skin.

I think Backend is more affected by AI than Frontend.

Front-end development is increasingly about UX Engineering—making decisions based on human psychology and accessibility laws. AI can’t feel frustration, so it doesn’t know when a UI pattern is frustrating.

AI has access to knowledge, but it cannot truly see skills in action, which is why it frequently hallucinates outcomes.