Grok：即使“编码者”使用大型语言模型，每天用C语言编写数千行代码也是一件大事吗？
自从我沉迷于“氛围编码”（即使用大型语言模型生成代码）以来，我几天前问了Grok，在像C这样的复杂语言中，每天生成数千行代码是否算多，尤其是这个项目涉及人工智能（所以对于复杂任务来说，代码行数是四位数/天，我们不是在谈论记事本克隆、销售点系统、牙科预约、加密钱包或任何初级开发者应该做的事情）。
不过，问题在于：虽然一个人可能是完全的新手，除了下载Python 3.9几乎不会编写代码，但如果你必须处理大量代码，你就得编译它，解决潜在的错误（在编译过程中），如果大型语言模型给出的代码错误地与目标相悖（例如：它自动在我的项目中添加了一个安全“对齐”功能，基本上禁止在我的计算机上运行“rm -rf”...但我在Windows上，所以我看到这个“安全”功能后就手动从代码中删除了它）。
问题是：初级开发者或刚入门的人和那些编程多年甚至几十年的人之间有什么区别？某种程度上，这就像在问“数学家使用计算器的方式与非数学家使用计算器的方式一样吗？”我想区别是微乎其微的？
查看原文
Since I got addicted to vibe-coding (to the unilluminated "vibecoding = using LLM to generate code), I asked grok couple of days ago if getting thousands or more LOC/day in complex language like say C is a lot or not, especially since the project involves AI (so it's 4-digit loc number/day for a complex task, we're not talking a Notepad clone, PoS, dental appointment, crypto wallet or anything that junior dev should do).<p>Here is the thing though: while one may be a total newbie and can barely type code apart from say downloading Python 3.9 if you have to deal with tremendous amount of code you'd have to compile, address potential errors (during compilation), if the LLM gives you code that erronously works against your goal (example: it automatically put a safety 'alignment' on my project basically forbidding 'rm -rf' to be run on my computer...but i'm on windows so i saw this "safety" feature and just manually deleted it from the code).<p>The question is: Is there any difference between a junior dev or rather someone just starting and someone who's been coding for years or even decades? In a way it's kind of like asking "do mathematicians use calculators the same way non-mathematicians use them?" I guess the difference is minimal?