Frontend Jobs Are Disappearing Everywhere: AI Collapses the Front/Back Split Into 'Full-Dry' Roles
The same logic that is eliminating frontend-specific roles in China will reshape hiring anywhere: AI compresses delivery times and makes role boundaries porous, so developers who can only do one side will face pressure to become requirement translators or deep specialists.
AI is erasing the standalone frontend job title, and the shift is moving from big tech to small and medium-sized companies. The argument uses Chinese employment statistics to show that SMEs are the main employers, yet still claims the frontend role is disappearing because AI has flattened the technical gap between frontend and backend work. If a frontend developer can describe backend requirements to AI, the backend specialist becomes redundant, and vice versa. The result is a revival of the 'full-stack' job description, with the new twist that the developer's main job is translating business needs into AI prompts rather than writing code by hand.
The piece also asserts that AI-written code is often more robust than what many humans produce, pointing to the common habit of skipping basic type checks. But it warns about a new risk: when deadlines shrink drastically, no one reads the AI-generated code. The process becomes requirements in, code out, and everyone from the customer to the tester ignores the implementation, turning the feature into a black box. This is the hidden cost of the speed gain.
The author sees two possible career paths: a deep specialist in a niche area like Canvas rendering or WebGL, or a generalist 'full-dry' engineer who can handle both sides. The post is a direct challenge to developers who still believe frontend positions will persist simply because SMEs are the largest employers.
The claim that frontend roles disappear even in SMEs rests on the assumption that AI capability automatically erases role boundaries, but it offers no hiring data from small companies to support that.
Using type checks as the main evidence for AI code robustness is narrow; production robustness also includes domain logic, error handling, and maintainability, where AI output may not consistently outperform experienced engineers.
The black-box risk is a real effect of compressed schedules, but it isn't unique to AI; human-written code becomes just as opaque when nobody reads it, AI just makes the delegation more complete.