738NGX completed a 6 month internship as a Software Engineer from July to December 2025. During his time here, he made significant contributions to RouteVN.

He was asked to write a reflection on his internship experience so that it can be helpful for future interns considering the position.

This reflection was originally written in Chinese. The English version was translated by AI. You can find the original Chinese version below.

In 2018, Sony released a game called Detroit: Become Human, which tells the story of three androids. I begin with this game because the difference between humans and androids is not merely biological; it is also about thought and behavior.

Today, the development of artificial intelligence is advancing rapidly, and we all find ourselves swept up in this transformation. Some learn to ride this wave and thrive, while others are swept away, becoming mere appendages to machines, human in biology alone.

My six-month internship at Yuusoft was an invaluable experience because it taught me far more than just work itself. More importantly, it taught me how to be an independent human.

Human progress has always been tied to the evolution and use of tools, but each iteration of tools has meant the elimination of many old professions. Historically, the reason for this elimination was often a subjective rejection of new tools.

But AI breaks this pattern: it threatens to eliminate not only those who reject it, but also those who mindlessly embrace it.

Initially, I reduced the job to a simple formula: feed requirements to an AI, verify the output works, and call it done. The problem is that this mechanical, procedural way of working fails to demonstrate human value. It reduces me to a middleman between supervisor and algorithm.

In this role, I became expendable, no longer human, but merely another tool in the chain, as replaceable as the android I resembled.

Over these months, I came to realize that AI's value lies less in productivity than in learning. It can help me quickly get up to speed on unfamiliar subjects, answer questions about half-formed ideas, and propose viable alternatives.

Beyond that lies human value: to examine, to critique, to decide. Tools can write code, but they cannot grasp the boundaries and trade-offs that shape it.

Only when we stop accepting code that merely "runs" and start asking "why does it run?" and "how long will it run?" do we transform from tool to toolmaker.

Androids exchange data; humans collaborate. The difference is everything.

Solo work may suffice in school or on personal projects, but it is a major pitfall in a work environment. By nature, I avoid socializing and prefer solving problems alone.

But this internship taught me to seek help without shame. Asking for help might feel like an imposition, but struggling alone, especially in the wrong direction, wastes far more time and jeopardizes the entire project.

Yuusoft's workflow embodies this through varied communication: asynchronous messages and document updates, synchronous standups and one-on-ones, code reviews and pair programming.

This constant communication felt uncomfortable at first. But I've learned it's not just efficient; it's what separates human teams from collections of isolated tools.

Androids execute instructions. They cannot question them. Reason doesn't factor in, only compliance.

I began as an executor: receive task, write code, submit. Done. I was accountable to the instruction, not the outcome.

But six months cultivated something different: Ownership. Ownership means looking beyond the task to examine the request itself. It means asking not just "does this code work?" but "is it robust? maintainable? Does it serve the user well?"

When problems arise, Ownership means we don't blame the tool or the instruction. We step up and solve them.

Ownership emerges when we adopt a holistic view: sacrificing speed for quality, polishing not just to finish but to excel. This is agency, the core of what makes us human.

These six months were more than an internship. They were an education in resistance: how not to be assimilated by machines.

In an era where AI technology is exploding exponentially, preserving our humanity has never been harder, or more essential. Technology keeps iterating, tools keep evolving, but those qualities that belong to humans (critical thinking, messy collaboration, ownership of outcomes) cannot be generated by algorithms.

This experience will anchor my career: a reminder not to go gentle into that good night, not to become an efficient but hollow tool. I must remain human, capable of pain, of thought, of warmth.

索尼在2018年发售了一个名为 Detroit: Become Human （底特律：化身为人）的游戏，讲述了围绕三个仿生人形机器人展开的故事。之所以举这个例子作为切入是因为，人和仿生人的区别不仅仅是生物意义上的，更是思想行为上的。

现如今人工智能技术的发展可以说是浩浩荡荡，每个人都身处洪流之中。其间有的人可以凭借自身的努力站在潮头之上风光无限，也有人就此被洪流所淹没，沦为了机器的附庸。虽然从生物学上说他们依然是人类，但实质上的行为已经与仿生人无二了。

在Yuusoft六个月的实习于我而言是一次宝贵的经历，因为它让我学到的绝不仅仅是工作本身，更重要的是让我学会做一个独立的"人"。

人类的发展离不开工具的演变和使用，但是每一次工具的迭代都意味着一大批旧的工种被淘汰。从历史来看，淘汰的原因往往是主观上对新工具的排斥；然而人工智能作为工具出现之后则打破了这一惯例：它将要淘汰的不仅仅是排斥它的人，还有滥用它的人。

就我而言，一开始理解的工作内容很简单：只需要把需求复述给AI工具，然后测试一下产出是否能正常工作就算完成一项验收了。

问题在于这种机械化、程序化的工作方式体现不出"人"的价值，只不过是在充当Supervisor和AI工具之间的沟通桥梁罢了，这样扮演的角色往往是无关紧要并且不被期待的，因为此刻人已不再是人，已经是和仿生人一样的工具了。

这几个月的工作下来让我意识到了AI相比于作为生产力工具，更重要的是作为学习工具在发挥作用。它能帮助我快速接触未上手过的新事物、解答一些在酝酿阶段的想法，提出一些可用的备选方案。

除此之外的事情呢？那就是体现"人"价值的地方了：去审视、去批判、去决策。工具可以帮我写出代码，但它无法理解代码背后的边界与代价。

只有当我们不再满足于代码"能跑"，而是开始深究"为何能跑"以及"能跑多久"时，我们才从工具的附庸，真正变成了工具的主人。

仿生人之间的沟通只需要简单的数据交换，但是人际之间的协作往往比这复杂。

或许在求学生涯或者完成个人项目时单打独斗是可行的，但是在工作环境中却是大忌。从我个人的性格出发来说，是不太喜欢社交并且倾向于只依靠自己解决问题的。

但是在这里实习的经历教会了我在工作中不要羞于向他人寻求帮助，即便看来是占用了别人的时间，但是靠自己解决问题甚至最后扎在了错误的方向上却是实打实浪费了更多的时间，从而会拖累整个项目进度。

现在Yuusoft的工作流中也有了多种不同的交流和写作方式。无论是异步的消息交流、共享文档更新、代码Review还是同步的周会、Standup、1on1、pair programming等，已经能极大打破成员之间的隔阂。

虽然可能刚接触时会不适应这种频繁快节奏的沟通，但是这对工作效率的提升却是必不可少的环节。

仿生人的程序设定决定了它们只能执行既定的指令，无论指令是否合理，它们都会一丝不苟地完成。

在开始的一段时间里，我把自己定位为一个"执行者"：接到任务，写代码，提交，结束。这种模式下，我只对"指令"负责，而不对"结果"负责。

但是长期的实习却是在培养人的Ownership，这意味着要跳出Task的框架，去审视Request本身。它要求我们不再满足于代码逻辑的正确，而是要去思考这段代码在整个系统中是否健壮、在未来维护中是否清晰、在用户体验上是否友好。

拥有Ownership就意味着当出现问题时，不是把责任推给工具或指令，而是主动站出来承担并解决。

只有当我们开始从全局的视角审视项目，愿意为了长远的质量而牺牲短期的速度，为了把事情做好而不仅仅是做完去反复打磨时，我们才真正拥有了独立意志。这是人之所以为人的核心——主观能动性。

回顾这六个月，它不仅仅是一次简单的实习，更是一场关于"如何不被机器同化"的教育。

在AI技术指数级爆炸的今天，保持"人类"的特质变得前所未有的困难也前所未有的重要。技术在不断迭代，工具在不断翻新，但那些属于人的特质——批判性的思考、充满摩擦的协作、以及对结果负责的担当——是永远无法被算法生成的。

这六个月的经历将成为我职业生涯的底色，提醒我无论何时，都不要温和地走进那个良夜，不要沦为高效但空洞的工具，而要始终保持痛感，保持思考，做一个有温度的"人"。