{
  "id": 19376436,
  "title": "Thoughts on AI Coding",
  "url": "https://urgent.news/2026/11/10/thoughts-on-ai-coding",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-10T18:27:33.000Z",
  "source": {
    "name": "Lobsters",
    "slug": "lobsters",
    "url": "https://danielbmarkham.com/blog/thoughts-on-ai-coding"
  },
  "original_language": "en",
  "account": "The rise of AI in coding presents a significant transformation for programmers, a job that has historically been largely unchanged. While the theoretical aspects of computer science remain vital, the real concern lies with practitioners - those who translate customer needs into functional computer systems. Traditionally, this meant any task involving instructing computers to accomplish useful work. However, the new definition incorporates tasks that customers have previously attempted without success. This often involves pre-generated code from large language models (LLMs) that highlight prior struggles.\n\nIn this evolving landscape, the new programmer may face the daunting task of maintaining and updating legacy systems. However, my hypothesis is that the future will see a surge in competition and misinformation, with individuals claiming to possess magical models or recipes to solve these problems. Ultimately, I believe we need to distinguish intent from execution. This concept, once emphasized in early coding practices, has since fallen out of favor. Defining intent is akin to creating a flowchart - outlining the general pattern for desired outcomes without getting bogged down in minute details. Execution, on the other hand, is about bringing that general plan to fruition, ensuring the results meet the expectations set forth in the flowchart.\n\nThe challenge lies in effectively communicating intent to AI models, as they excel at identifying patterns but struggle with nuanced understanding. My analysis suggests that separating intent from execution could streamline the coding process. However, a critical question remains: could AI models become so sophisticated that they could effectively interpret and cover all possible error patterns humans make when expressing intent? While ongoing research aims to address this, the future of programming undoubtedly hinges on our ability to clearly articulate intent and trust AI to execute it accurately. Despite the rampant hype and noise from online experts, value and progress will prevail for those who prioritize clarity and precision in their communication.",
  "summary": "This is a commentary on the future of programming as a job. It is written by someone who grew up alongside the rise of personal computers. Comments",
  "key_points": [],
  "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."
}