编码会因为人工智能接管而消亡吗?
Is coding dead because AI has taken over it?

原始链接: https://www.jehuamanna.com/blog/2026/is-coding-dead/

就像新的交通方式从未否定学习步行的重要性一样,人工智能也不应掩盖学习基本编程技能的必要性。步行是一种复杂的身体过程,需要调动许多肌肉和神经系统——需要练习和发展。同样,编程不仅仅是关于输出;它是一种严格的智力锻炼,激活大脑中负责逻辑、记忆、解决问题和创造力的关键区域。 人工智能是加速开发的强大*工具*,但它不能取代有效软件设计和解决问题所需的 core 思维能力。仅仅关注人工智能生成的代码(“Vibe Coding”)而缺乏计算机科学原理的坚实基础,就像放弃步行练习而只选择乘车一样——限制了基本能力并阻碍了真正的创新。真正的进步需要同时利用工具*和*培养潜在技能。

## 编码已死?Hacker News 讨论 最近 Hacker News 论坛讨论了人工智能是否预示着编码职业的终结。 结论倾向于**否,但形势正在变化。** 许多评论者认为软件开发工作不会消失,但*手动编写的代码量*可能会减少。 人工智能,特别是 LLM,被视为强大的 RAD(快速应用开发)工具,使团队能够承担更雄心勃勃的项目,并可能提高开发人员的产出期望。 一位开发人员分享了过去由于开发周期缓慢而阻碍盈利的经历,暗示人工智能可能有所帮助。 然而,人们对过度依赖人工智能表示担忧。 虽然人工智能在自动化任务和提供高级解决方案方面很有帮助,但它缺乏避免错误的细致判断力——就像给没有接受过适当培训的人一把链锯。 一些人指出编码仍然令人愉快且对于构建强大的系统至关重要,而另一些人则指出人工智能的能力可能会严重影响网络开发等特定领域。 最终,这场讨论强调的是编码的转型,而不是灭绝。
相关文章

原文

Influential personalities propagate such conclusions.

When different modes of transportation appeared over time in the past, it always accelerated the quality of mobility. This includes ships, trains, motorbikes, cars, and even airplanes.

During that time of transformation, none of the parents discouraged their child from learning to walk. Isn’t it? They never said, “Listen, child, walking is dead; learning to walk is of no use, because in the future we will have all these modes of transportation to get you wherever you need to go.”

The parents did not say so.

Why? Walking is a fundamental and necessary skill that will always hold importance, irrespective of technological advancements.

Walking is a learned behavior. You must go to get the vehicle, regardless of what you chose.

But you might say that I will minimize the number of steps while walking and use the mode of transportation efficiently.

  • Do you know that to take one footstep, you need 30–40 major muscles actively involved normally? This includes muscles from both legs, the pelvis, the core, and even the arms.

  • At minimum, about 10–15 major joints are continuously coordinated.

  • Walking uses a set of nervous systems at the same time:

    • The spinal cord serves as an automatic rhythm generator.
    • The Brainstem & cerebellum serves to coordinate and balance.
    • The Cortex manages conscious control and adaptation.
    • Sensory systems provides constant feedback.

Back to programming, while doing so, programming activates a network, mainly:

  • Prefrontal cortex (DLPFC) – working memory, planning, control (most important)
  • Parietal cortex – logic, math, symbol manipulation
  • Temporal lobes – memory and meaning of code
  • Inferior frontal gyrus (language areas) – syntax and rules
  • Visual cortex – reading and scanning code
  • Anterior cingulate cortex – error detection and debugging
  • Cerebellum – fluency and automation
  • Default Mode Network – creativity and “aha” insights

Thus it mainly involves using short-term memory, decision-making, and logical thinking, guided by the prefrontal and parietal cortex, with help from memory, vision, and error-checking systems.

Just like walking requires a lot of physical effort for moving, which is learned and should be practiced, programming requires a lot of mental effort for thinking, which is learned and should be practiced.

On the other hand, AI is merely a tool. AI serves to swiftly process and transfer our thoughts. But they have never been the replacement for thinking, creativity, and imagination.

So do you think it is ok to tell people to focus just on AI-generated code (aka: Vibe Coding) rather than learning and practicing Computer Science fundamentals, programming paradigms, design patterns, and the ability to design simple and complex software systems?

联系我们 contact @ memedata.com