Overview

Martin Fowler observes that LLMs are fundamentally changing software development roles by making platform-specific expertise less valuable than AI-driving skills. This raises questions about whether the industry will embrace generalist developers or continue working around existing silos.

View Original

Key Arguments

  • LLMs are eating specialty skills - traditional front-end and back-end specializations are becoming less important: As AI tools become more capable of generating code across platforms, the specific technical knowledge of particular frameworks or languages matters less than the ability to effectively direct and work with LLMs
  • LLM-driving skills are becoming more valuable than platform-specific expertise: The ability to effectively prompt, guide, and collaborate with AI systems is emerging as a more critical skill than deep knowledge of specific technologies

Implications

Software developers need to adapt their career focus from deep specialization to AI collaboration skills. Organizations must decide whether to restructure around generalist “Expert Generalist” roles or continue with siloed teams that use AI tools. The fundamental question is whether AI will flatten the skill hierarchy in software development or simply change the tools used within existing structures.

Counterpoints

  • LLMs may reinforce silos rather than eliminate them: Instead of creating more generalist developers, AI tools might simply enable developers to code around existing organizational and technical silos without fundamentally changing how teams are structured
  • Deep technical expertise may remain necessary: Complex systems, performance optimization, and architectural decisions may still require specialized knowledge that LLMs cannot fully replace