{
  "id": 10483171,
  "title": "The problem is not the AI code, but nobody knows anything anymore",
  "url": "https://urgent.news/2026/09/28/the-problem-is-not-the-ai-code-but-nobody-knows-anything-anymore",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-28T16:11:42.000Z",
  "source": {
    "name": "Hacker News",
    "slug": "hacker-news",
    "url": "https://www.ssp.sh/brain/the-problem-is-not-the-ai-code-but-nobody-knows-anything-anymore/"
  },
  "original_language": "en",
  "account": "The crux of the issue at hand is not the AI code itself, but rather the widespread lack of understanding among individuals and teams regarding system architecture and the underlying intent behind coding decisions. Even if AI-generated code is of average quality, depending on the task and size, it has the capability to raise the standard to an acceptable level.\n\nAccording to the wire material, the problem lies in people not having any knowledge at all, and merely relying on AI platforms like Claude to provide answers. This has led to a complete lack of planning and organization, which is evident in fast-moving startups, larger companies, and scenarios where middle management is heavily pushing AI implementation. The sentiment expressed by the author is that engineering has reached a state that is far from ideal, as exemplified by their recent start at a major company where team members feel overwhelmed and overworked, often working 12 to 13 hours a day to push out code.\n\nThe lack of knowledge and understanding has led to a sense of helplessness among engineers, who are discouraged from reading and comprehending the codebase or the systems they are working on. This has resulted in an absence of any sense of accomplishment or progress, as everyone is essentially relying on LLMs (Large Language Models) to do their work. This trend is particularly acute in data engineering, where the author notes that \"people who grew up pre-AI\" had to possess extensive knowledge about the product and business to navigate the domain, but AI has now rendered that knowledge obsolete.\n\nThe author acknowledges that data professionals who grew up in the pre-AI era had to possess deep knowledge about the product and business in order to navigate the domain effectively. However, the advent of AI has made this knowledge obsolete, leaving newcomers in a state of confusion and uncertainty. The implications of this are far-reaching, as the ability to code and design with intent and architectural awareness have become increasingly important in today's software engineering landscape.\n\nIn conclusion, the ultimate challenge remains maintainability. As AI continues to generate code rapidly, the burden of maintaining the resulting systems and pipelines becomes increasingly heavy. Without a solid understanding of the underlying architecture and intent, it becomes exceedingly difficult to maintain and scale these systems effectively. While AI can undoubtedly assist in generating code quickly, the human element remains crucial in directing and orchestrating these efforts, emphasizing the importance of intent, taste, design, and architecture in today's software engineering world.",
  "summary": null,
  "key_points": [
    "Lack of understanding in system architecture and coding intent",
    "Relying on AI platforms like Claude for answers without planning",
    "Data professionals' knowledge made obsolete by AI advancements"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}