人工智能时代的初级开发者
Junior Developers in the Age of AI

原始链接: https://thoughtfuleng.substack.com/p/junior-developers-in-the-age-of-ai

## 意想不到的初级工程师危机 尽管过去曾担心开发者短缺,但现在却出现了一种令人惊讶的现象:大量受过训练的初级软件工程师面临艰难的就业市场。这不仅仅是速度放缓——即使行业在增长,对初级职位的需求也在*减少*,而对高级工程师的需求却在上升。 人工智能经常被认为是原因,但它更像是一种加速剂,加速了对经验丰富人才的现有偏好。核心问题不仅仅是*编码*能力,因为人工智能现在可以复制这种能力,而是*工程*——理解和维护复杂的、不断发展的系统,需要机构知识。初级工程师对于培养未来的高级专业知识、确保系统的长期健康至关重要。 优先考虑短期人工智能收益而忽视对初级人才的投资,是公司犯下的一个关键错误。招聘初级工程师不仅仅是为了填补职位;而是为了建立有韧性的组织,促进创新,并利用年轻一代独特的活力和人工智能采用率。 具有前瞻性的领导者认识到,强大的初级工程师人才储备是一项高杠杆的投资,尤其是在人工智能驱动的世界中,并且发现借助人工智能的辅助,现在入职培训更快、更容易。

## 初级开发者与人工智能转变 - Hacker News 摘要 一篇 Substack 文章引发了 Hacker News 的讨论,探讨了人工智能对初级开发者就业市场的影响。一个关键点是,**经验丰富的工程师优先编写简洁、易于维护的代码**,这项技能对于应对系统复杂故障时的扩展支持至关重要——这是人工智能无法完全取代的。 许多评论员认为,当前的转变并非*仅仅*由于人工智能,而是多种因素共同作用的结果:ZIRP 结束、编码职业的“迷因化”、远程工作的兴起(以及面试作弊),以及由此导致的申请者激增,推高了高级工程师的薪资。这使得公司更倾向于招聘有经验的员工。 然而,也有人认为,**计算机科学毕业生人数过多**,加上强大的人工智能工具,形成了一批*价格低廉*但实力不俗的初级人才。建议是,公司应侧重于现场面试,并专门招聘初级人才以获取成本效益高的劳动力。 争论的焦点还包括评估初级开发者的技能,一些人提倡询问关于“编码代理”的问题,以衡量他们对当前行业趋势的了解。
相关文章

原文

For a long time, we were all hand-wringing over the shortage of software developers. School districts rolled out coding curriculums. Colleges debuted software “labs”. “Bootcamps” became a $700m industry.

Today, we have the opposite problem. Thousands of trained, entry-level engineers that no one wants to hire.

Just as software finished eating the world, zero interest rates ended. Companies optimized for cash and slowed hiring. The market didn’t shrink, but stopped growing at the breakneck pace we all expected.

The result: a glut of entry level talent groomed for jobs that never materialized.

This would explain a more competitive entry level market. But it doesn’t explain the entry-level market shrinking, despite overall industry growth.

In short: demand for senior talent is rising, but has fallen off a cliff for juniors.

AI didn’t create this trend - there was already a bias for senior talent pre-2022 - but it gave leaders a convenient justification to exacerbate it.

In terms of speed, price, and quantity, juniors can’t compete with LLMs. Code is now a commodity.

It sounds logical because it’s true, but it misses the big picture.

Here’s the core misunderstanding:

The job of software engineering is not “writing code”.

Coding: translating a process into something a machine can understand and execute

Engineering: sustaining and evolving interconnected, ever-changing systems

When we talk about “the code” we often actually mean the systems that contain code.

These systems are complex. They include layers of interwoven dependencies that evolve unpredictably. Between unique technology choices, historical decisions, and company-specific processes, every system’s “code” is its own special snowflake.

Fail to understand your special snowflake, and things break. Ex: If Apple makes an update, you can’t respond if you don’t know how your mobile app connects to your billing system.

It’s why the best engineers try to write as little code as possible. They understand that each additional line is another thing that must be understood, remembered, and maintained.

That understanding, or institutional knowledge, lives in human minds.

And humans quit. They move. They retire. They die.

This has been and remains the primary function of juniors: to ensure an evergreen supply of future seniors.

Until AI solves human mortality, this will not change.

In fact, for “AI-first” companies, this is doubly true. Absent expertise and institutional knowledge, AI is just expensive (and potentially destructive!) autocomplete.

Companies who have stalled hiring to “wait out” the full impact of AI are prioritizing a hypothetical optimization problem over an impending inevitability.

Not to mention, it screws over an entire generation.

Gen-Z doesn’t know life before the internet. The pandemic stole their formative social and professional years.

To their colleagues - with a shared understanding of How Things Are™ and Back In My Day™ - they’re practically feral.

There’s some scientific truth to this: 20-somethings are inherently narcissistic. Wisdom requires having a full frontal lobe.

But this is what functional societies do. We lament how uniquely terrible the youth are… then teach them anyway.

AI has changed a lot, but it hasn’t eliminated our responsibility to our young people. Leaders who don’t prioritize the next generation are failing to pay forward what was given to them.

Most leaders aren’t trying to undermine over society or create existential risks to business continuity.

But when timelines are tight, budgets are tighter, and things need to get done… “hiring juniors” quickly becomes a nice to have.

Reframe the problem: you’re not “trying to hire juniors”, you’re laying the infrastructure for a fundamentally stronger organization.

An engineering org that “can’t afford” junior talent is incredibly fragile.

Teams that can easily absorb junior talent have systems of resilience to minimize the impact of their mistakes. An intern can’t take down production because no individual engineer could take down production!

This doesn’t mean you need unit tests for every edge case or rock-solid infrastructure. It can be as simple as a formal mentorship program… or implementing common-sense policies like “don’t deploy on Friday”.

These guardrails and programs aren’t “for juniors” - all engineers are benefitting from a system that promotes growth, experimentation, and learning. AKA: The foundation of innovation.

When you get this right, hiring entry-level engineers feels natural and obvious.

That feral hunger and drive (only healthy when you don’t have a full-frontal lobe) is a powerful force. While writing this, I spoke with 24+ technical leaders. Every single one could name an exceptional junior engineer they were mentoring.

"Some of our top engineers at Finley were all entry level hires,” says Kevin Suh, CTO at Finley, “It’s impossible to replicate that energy, internal motivation, and resourcefulness."

This isn’t “grind” or “grit” or “fire” - it’s the miracle of early learning.

The magic of those working toward true mastery is inspiring and necessary in its own right, but proximity to the high-growth only juniors can experience is infectious and energizing.

Perhaps the most compelling reason to hire juniors? AI itself.

The biggest barrier to AI transformation has been workforce buy-in. Gen Z not only leads in AI adoption, but acts as an accelerant: nearly two-thirds of Gen Z workers help their older colleagues learn AI tools and workflows.

At the same time, AI has also brought down the cost of onboarding.

“Our top juniors are now getting up to speed in 3 months,” says Arjun Kannan, CTO of Residesk, “AI has made coaching much easier because the kids are learning the basics on their own.”

If you already have the infrastructure, hiring juniors may be one of the highest-leverage AI bets available today.

“Your greatest alpha as a hiring manager is finding people who are young, smart, and unproven,” says Randy Brown, CTO of Scout, “and that’s outrageously easy to do right now.”

If you can see through the hype, you gain a huge advantage. Not only do you get top talent in a favorable market, you get an injection of eager, learning energy that can propel your entire team into the future.

That future will be shaped by AI, but AI can’t change everything. Systems - and society - still require care and renewal to endure.

You remembered that spring would come again… and faithfully planted seeds.

A huge thank you to the small village that helped read drafts, share quotes, and provide feedback on this article: Kathryn Minshew, Randy Brown, Kevin Suh, Arjun Kannan, Matthew Casey, Charity Majors, Sadam Iqbal, Maria Ashby, @bentloy, and Bea Arthur.

About the Author:

Christine Miao is the creator of technical accounting–the practice of tracking engineering maintenance, resourcing, and architecture. It visualizes the most complex technical problems - think: breaking up monoliths or cleaning up tech debt - in a way that anyone can understand.

Learn More

联系我们 contact @ memedata.com