五层级:从辣味自动补全到软件工厂
The Five Levels: From spicy autocomplete to the dark factory

原始链接: https://www.danshapiro.com/blog/2026/01/the-five-levels-from-spicy-autocomplete-to-the-software-factory/

代码成本的快速下降——“技术通缩”——要求我们改变处理技术债务的方式。聪明的团队正在战略性地推迟人工编码时间,预计未来会有更便宜的AI驱动解决方案。然而,仅仅将AI用于小的任务,如正则表达式,不足以利用这种通缩。 作者概述了五个层次的AI代码自动化,类似于美国国家公路交通安全管理局(NHTSA)的驾驶自动化等级。**0级**是纯手动编码,AI仅作为搜索辅助。**1级**卸载离散的任务(单元测试、文档字符串),但保持类似的工作流程。**2级**(目前大多数处于此级别)涉及与AI配对以提高生产力。**3级**将开发者转变为管理者,主要审查AI生成的代码。**4级**将角色转移到产品管理,专注于规格和规划,而AI处理执行。最后,**5级**代表完全自动化的“黑工厂”,人类干预最小。 作者目前处于4级,强调了AI持续进步带来的彻底变革的潜力——以及陷入之前不喜欢角色的风险。

## “黑暗工厂”与AI生成代码 – 摘要 最近Hacker News上出现了一场关于“黑暗工厂”的讨论——即AI自主生成代码且无需人工审查的系统。一位参与者描述了一个小团队成功地通过这种方式构建出令人信服的软件,他们专注于强大的测试和系统设计,而让AI处理编码本身。 然而,怀疑论普遍存在。许多评论者质疑这种方法的长期可行性,预测会出现一个快速炒作周期,随后是对其局限性的现实评估。担忧包括潜在的低质量、难以维护的代码,以及缺乏人为监督对安全性和可靠性的保障。 一个关键的争论点在于AI是否会*取代*开发者,还是仅仅*增强*他们。一些人认为AI将在生成一次性代码以完成快速任务方面表现出色,而另一些人则强调在规范、设计和长期维护方面持续需要人类专业知识。有人认为,价值将从代码本身转移到定义良好的规范。 最终,这场讨论凸显了AI编码的快速进步与构建真正有价值且可持续软件的持久挑战之间的紧张关系。许多人认为,当前一波AI生成项目更多的是关于实验和“数字兴奋剂”,而不是创造持久的解决方案。
相关文章

原文

In my last post, I wrote about technical deflation. We’re seeing the cost of code is dropping so fast that we need to change our tech debt payment plans. The smart teams are deferring payment on human hours today to pay them back with cheaper AI hours tomorrow.

But how do you actually cash in on those cheap hours?

If you are just using ChatGPT to write your regex, you aren’t really getting the benefits of deflation. You’re just typing faster. 

I’ve now seen dozens of companies struggling to put AI to work writing code, and each one has moved through five clear tiers of automation. That felt familiar, and I realized that the federal government had been there first – but for cars. 

In 2013, the NHTSA created the five levels of driving automation1. This was helpful, because while the highest level at the time was only level 22, it let everyone have a common language for both where things were, and where things were going.  

Level Zero is your parents’ Volvo, maybe with an automatic transmission. Whether it’s vi or Visual Studio, not a character hits the disk without your approval. You might use AI as a search engine on steroids or occasionally hit tab to accept a suggestion, but the code is unmistakably yours. This is manual labor in a deflationary world.

At Level 1, you’ve got lanekeeping and cruise control. You’re writing the important stuff, but you offload specific, discrete tasks to your AI intern. “Write a unit test for this.” “Add a docstring.” You could be using anything from copy-paste ChatGPT to Copilot. You’re seeing speedups, but your job is unchanged. You’re still moving at the rate you type.

At Level 2, you’ve got Autopilot on the highway. As a coder, you feel free. You’ve got a junior buddy to hand off all your boring stuff to. This is where 90% of “AI-native” developers are living right now. You are pairing with the AI like a colleague. You get into a flow state; you’re more productive than you’ve ever been. You’re not using chat, you’re getting real mileage out of an AI-native coding tool. But here is the danger: level 2, and every level after it, feels like you are done. But you are not done.

Level 3 is a Waymo with a safety driver. You’re not a senior developer anymore; that’s your AI’s job. You are… a manager. You are the human in the loop. Your coding agent is always running multiple tabs. You spend your days reviewing code. So much code. Your life is diffs. For many people, this feels like things got worse.

And almost everyone tops out here.

Level 4 is a robotaxi, and while it’s driving, you can do something else. You’re not a developer. You’re not a development manager either. You’ve now become that which you loathed: you’re a PM3. You write a spec. You argue with it about the spec. You craft skills (for Claude Code, because most folks at level 4 seem to find their way to Claude Code). You plan schedules. You review plans. Then you leave for 12 hours, and check to see if the tests pass.

I’m here.

At level 5, it’s not really a car any more. You’re not really running anybody else’s software any more. And your software process isn’t really a software process any more. It’s a black box that turns specs into software.

Why Dark? Maybe you’ve heard of the Fanuc Dark Factory, the robot factory staffed by robots. It’s dark, because it’s a place where humans are neither needed nor welcome.

I know a handful of people who are doing this. They’re small teams, less than five people. And what they’re doing is nearly unbelievable – and it will likely be our future.

Thanks to Jesse Vincent, Justin Massa, Ramon Marc, and Noah Radford (who wrote this incredible piece) for reading drafts of this.

You might want to subscribe or follow me on Twitter so you don’t miss new articles.

联系我们 contact @ memedata.com