一名初级软件工程师在探讨AI代理时代的职业发展时，总结了资深工程师的关键能力特征1。这些能力包括为已发布代码进行辩护、论证设计选择、识别AI错误方向、判断何时委派工作何时亲自动手，以及花时间指导他人1。作者指出，好奇心、代码审查、自主性和协作这四项核心技能无法通过AI获得，必须通过主动学习和实践来培养1。
为了在AI时代避免成为"永久初级"，作者采取了一系列具体行动1。他通过编写skill.md来约束AI行为，其中明确规定AI不能直接编辑文件、运行迁移或提交代码，直到获得明确确认1。同时，作者坚持偶尔手写部分代码以测试理解、学习架构设计、掌握算法原理、提高token效率并增强信心1。此外，他还主动向资深工程师请教，学习代码评审模式1。这些努力最终取得成效——作者从最初频繁打开skill.md的状态，逐步降至每天打开一次，并最终获得了提前晋升的提名1。
A junior software engineer has identified four core competencies that artificial intelligence cannot cultivate, arguing that technical professionals must actively develop these capabilities to advance beyond entry-level positions in an AI-augmented workplace.1 The engineer emphasizes that senior developers distinguish themselves by defending deployed code, justifying design decisions, recognizing when AI suggestions lead in wrong directions, determining when to delegate versus execute personally, and investing time in mentoring others.1
The four irreplaceable skills are curiosity—consistently asking questions to deepen understanding; code review proficiency—learning established patterns and standards through reviewing others' work; autonomy—periodically writing code by hand rather than relying entirely on AI assistance; and collaboration—seeking guidance and knowledge from experienced engineers.1 Writing code manually serves multiple purposes: validating comprehension of concepts, learning architectural design patterns, understanding algorithmic principles, improving token efficiency in AI interactions, and building professional confidence.1 The engineer implemented a skill.md file that constrains AI behavior by preventing the tool from directly modifying files, executing migrations, or committing code without explicit approval.1
This deliberate approach to skill-building yielded measurable results: the engineer's frequency of consulting the skill.md file decreased from checking it with every task to once daily, and ultimately received an early promotion nomination.1 The case demonstrates that junior engineers can avoid remaining perpetually junior by combining AI tools strategically with intentional, self-directed learning and mentorship-seeking rather than outsourcing all technical decision-making to automation.1
评论
还没有评论，欢迎留下第一条。