It matters who teaches you

师承何人，至关重要

Falling in love with the machine

爱上机器

I struggled throughout my time at school. I always loved learning, and still do, but the structure and environment of school never really worked for me. So now, at 37, watching the rise of tools where you type a few sentences and an LLM spits out a hundred different explanations of a topic, I naturally fell in love. 我在学生时代一直过得很挣扎。我一直热爱学习，至今未变，但学校的体制和环境从未真正适合我。所以现在，37岁的我看着这些工具的兴起——你只需输入几句话，大语言模型（LLM）就能吐出关于某个主题的一百种不同解释——我自然而然地爱上了它们。

Even though their answers are statistical predictions, with good inference they more often than not come close to a real, “true” answer. And even when they don’t, bouncing ideas back and forth has already unlocked insights I would never have had otherwise - especially as a remote worker sitting at home. I wonder what would have happened to my academic career if these tools had been available back then. I could have taken notes in class, gone home, and spent hours trying to really understand a topic I was too shy to ask follow-up questions about, or one I had missed because I wasn’t paying attention. 尽管它们的回答是基于统计预测的，但通过良好的推理，它们往往能接近真实、“正确”的答案。即使有时不然，这种思想的碰撞也已经为我开启了原本无法获得的洞察力——尤其是作为一个居家办公的远程工作者。我常想，如果当年有这些工具，我的学术生涯会是怎样？我可以在课堂上做笔记，回家后花几个小时去真正理解那些我因害羞而不敢追问、或因走神而错过的知识点。

This thought sat with me for a long time, and it shifted my overall position towards seeing LLMs as a net positive (provided we get the regulatory environment for building data centers, training, and usage in order). 这个想法在我心中盘旋了很久，它改变了我对大语言模型的整体看法，使我将其视为一种净收益（前提是我们能理顺数据中心建设、模型训练和使用方面的监管环境）。

A doorway many others have passed through

许多人曾走过的门径

That is, until I read the book “Inexplicable Magic” by Susan Piver. Now, this book has nothing to do with LLMs or even programming in general. It’s a book about meditation and Buddhism. 直到我读了苏珊·皮弗（Susan Piver）的《不可思议的魔法》（Inexplicable Magic）。这本书与大语言模型甚至编程毫无关系，它是一本关于冥想和佛教的书。

In it, she touches on something I was subconsciously aware of but could never find the right words for. She writes: 书中，她触及了我潜意识里有所察觉却一直无法用言语表达的东西。她写道：

Who you learn from is important because you have entered the practice via a certain doorway through which many others have passed. This doesn’t mean you can’t walk out and go through other doorways. It’s simply a factor that bears noticing. “你向谁学习很重要，因为你是通过某扇许多人曾走过的门径进入修行之道的。这并不意味着你不能走出来去尝试其他门径，但这确实是一个值得注意的因素。”

And further on: 随后她又写道：

[…] Is there a lineage behind your instructor and their instructions? If so, good, because we don’t want any made up bullshit. “[…] 你的导师及其教导背后是否有传承？如果有，那很好，因为我们不需要任何胡编乱造的废话。”

Using an LLM is exactly that. There is no lineage behind it and nobody who stands behind the answer, it’s made up bullshit that sounds convincing. You should learn from people who are free of bullshit. 使用大语言模型正是如此。它背后没有传承，也没有人对答案负责，那只是听起来很有说服力的胡编乱造。你应该向那些不讲废话的人学习。

The people who shaped me

塑造我的人

I am currently writing the second edition of my book “Rust Web Development”. When I approached Manning with the idea, I wondered why I would even write a book in this day and age. While writing the first edition, I kept thinking of all the teachers, senior engineers, and managers who shaped me into the professional I am today. They contributed as much to my growth and to who I am as all the work I put into my career myself. 我目前正在撰写《Rust Web开发》一书的第二版。当我向Manning出版社提出这个想法时，我曾自问：在这个时代，为什么还要写书？在写第一版时，我不断想起那些塑造了今日之我的老师、资深工程师和经理们。他们对我的成长和个人塑造的贡献，丝毫不亚于我在职业生涯中所付出的努力。

Their guidance, their wisdom, their view of the industry and my profession (and, to be honest, what kind of humans they were) still shape my worldview today. Later in the book, Susan Piver suggests holding someone in mind while you meditate, as a way to give back and connect with the people who inspire you and mean something to you. 他们的指导、智慧、对行业和职业的见解（坦白说，还有他们作为人的品格），至今仍在塑造我的世界观。在书的后半部分，苏珊·皮弗建议在冥想时心中默念某人，以此作为一种回馈，并与那些激励你、对你意义重大的人建立连接。

Before starting my professional software engineering career, I did an apprenticeship as an electrician, and it wasn’t even a question that you had a foreman who guided and taught you. Depending on your skills and aspirations, you would (ideally) be paired with the right foreman, one who could challenge you and grow you in the way you wanted and needed. This person looked you in the eye, had a gut feeling and experience, and decided what you were ready for - and capable of. 在开始职业软件工程生涯之前，我曾做过电工学徒。那时，你有一位指导和教导你的工头是理所当然的事。根据你的技能和志向，你（理想情况下）会被分配给合适的工头，他能挑战你，并以你想要和需要的方式让你成长。这个人会注视着你的眼睛，凭借直觉和经验，判断你准备好了什么，以及你能胜任什么。

I still think fondly of all the guidance, mentorship, and teaching I received. Every day, at work and in my private life, their teachings and work are part of my own. 我至今仍深情地怀念我所获得的所有指导、教诲和教导。每一天，无论是在工作中还是在私生活中，他们的教诲和工作方式都已成为我自身的一部分。

What gets lost

丢失了什么

What will all of this look like if we let smart-sounding machines replace real human connection and teaching? I love having these LLMs at my fingertips, but after two or three hours of bouncing ideas back and forth, I more often than not feel empty and burned out. When I pair program, or sit in a room with other people trying to find a solution, I almost always come out of those long sessions inspired and motivated. I often keep working after my normal workday is done, because I still feel the fire in me, the urge to create. 如果我们让听起来聪明的机器取代真正的人际连接和教学，这一切会变成什么样？我喜欢触手可及的大语言模型，但在与它们进行两三个小时的思想碰撞后，我往往感到空虚和精疲力竭。而当我进行结对编程，或与他人共处一室寻找解决方案时，我几乎总能在这些漫长的会议后感到备受鼓舞和充满动力。我经常在正常工作时间结束后继续工作，因为我依然能感受到内心的火焰和创造的冲动。

An LLM has no sense of who you are as a person. It interprets your text and spits out paragraphs that sound right. There is no teaching, no connection. If anything, I am writing this as a reminder to myself that nothing will ever replace the emotional connection we build with other humans, and the teaching that comes with it. We carry their thoughts, ideas, and outlook with us, let them shape us, and carry them forward to the next generation. I don’t think I will ever tell my children tales of my long coding sessions with an LLM, but I will tell them about the long nights and intense sessions I had with other humans: moments that shaped and defined who I am today. 大语言模型并不了解你是一个怎样的人。它只是解读你的文字，然后吐出听起来正确的段落。这里没有教学，没有连接。如果说我写这篇文章有什么目的，那就是提醒自己：没有什么能取代我们与他人建立的情感连接，以及随之而来的教导。我们承载着他们的思想、理念和观点，让他们塑造我们，并将这些传承给下一代。我想我永远不会向我的孩子们讲述我与大语言模型进行漫长编程会话的故事，但我会告诉他们那些我与他人共度的漫漫长夜和紧张时刻：正是这些时刻塑造并定义了今天的我。

Final thoughts

结语

These thoughts give me enough motivation to keep writing books. Even though I know they won’t have a massive impact on society, I hope that at least one young engineer will read one, look at the cover, and someday tell someone that they learned this from Bastian Gruber, and not from a statistically generated output produced many years ago. 这些想法给了我足够的动力去继续写书。尽管我知道它们不会对社会产生巨大的影响，但我希望至少有一位年轻的工程师在读到它时，看着封面，并在某一天告诉别人：他是从Bastian Gruber那里学到的，而不是从多年前生成的统计输出中习得的。