大部分人都不懂信息技术

Don't learn computer

Posted by Zeusro on November 10, 2025

最近通过DNS发现,这个网站的域名,海外流量超过国内了。 前段时间回高中,发现学文科的人好少。大部分人都挤进了理科,特别是去学计算机。

大部分人都不应该盲目学计算机

但在我看来,包括我在内,其实大部分人都不太适合学计算机——我只是阴差阳错选了理科,然后被迫在计算机这条路上一条路走到黑。

在一开始从业的时候,我积极认真学习,一边认真工作。但在三线城市,其实没啥意义,努力总有天花板和上限。而三线城市的上限就是自己的父母。

虽然在大部分人看来,我的天资不是一般的高,因此羡慕妒忌恨我的人无数。但在我看来,我写的代码其实很少,“缺少一个大项目”,我的能力的上限是我的时间,而且我已经开始感受到身体的衰老。

入行10年,发呆的时间其实比工作多。而且由于国内这个有毒的环境,很少能得到重用。一开始我鄙视周围那些学历低,还故意在我身边吸烟的狗叼同事;而等到我来到大城市,才发现自己成为被鄙视的那位。

国内环境缺乏对人的基本尊重,总是强调各种“控制理论”——未成年时通过权威关系绑架你(父母老师一定对,你们一定要听我的); 毕业之后通过工作限定你; 等你钱挣差不多了,就可以“买房结婚生小孩”全流程套餐来一遍。买一套房不够,还得给嫡长子买一套,再给孙子买一套。

然而,物质资源是有限的,互联网的岗位也是有限的。每年高考毕业生有几百上千万人,如果他们都贪图这个行业的“高薪”而进入这个赛道,那么无效内卷就会迫使这个行业变得低薪。

当上班变成一种表演,尽管没什么事儿,每个人还要在工位上假装上班。我在大公司边缘OB了几年,基本已经对他们完全祛魅。

互联网行业的悖论

更何况,互联网行业的第一悖论是“Kill Everything”——越优秀的从业者,必定越反社会

互联网行业的底层根基是信息技术,而信息技术代表的是一种新质生产力,它一开始的设定是“辅助人类”,但是渐渐地变成“取代人类的工作”,像是开源项目吞噬闭源软件公司,Kubernetes 取代了人力运维,AI毁灭了基层编程工作(中专学计算机还不如学修车)。

在中国,基本上所有互联网的业务都已经开发殆尽,但互联网企业却没有几家——大家普遍缺乏一种互联互通的精神,总是想着闭门造车,造一个“大平台”让其他人加入,然后自己作为“地主”收租。

但现在AI出来后,我们有一次“拨乱反正”的机会。在制定AI伦理相关法律的时候,我觉得“以人为本,以机械为辅”必定要放在第一位。

而且计算机相关的教材,一定要开始重写。

制定教材的人根本不懂计算机

Python语言中,通过缩进表示嵌套层次。如果按照既有思路,用纸质课本去印刷,那么缩进变成空白,其实可读性更低了。

更别提国内的教材滥竽充数,质量真是一言难尽。 过期的语言书籍,长期不下架,行文的作者思想迂腐,不懂变动。 他们写出来的内容跟天书一样难懂,只让人上课想睡觉。

这些作者最大的问题在于“懒惰”。他们认为自己混了那么多年,终于变成教育界权威人士,可以决定教材内容了,思想就开始变得懒惰,不再拥抱新的变化。

而另一方面,语言标准本身也在变化。虽然远古时期的 C语言C++依旧在发光发热,但这种晦涩难懂的语言显然不适合大部分初学者。

C#语言虽然优美,但缺乏生态链;Java语言自称跨平台,但在云原生领域如同小丑(Java in docker);Go语言号称简单,但在泛型和集合操作方面,远不如10年前,C#linq + lambda 那样优雅。

而计算机在我们发展了30多年,但我们唯一的编程语言是“内斗语言”。 我们只有在内部消耗这个问题上取得了空前一致的团结。

那些学计算机的人才,混了个大学教职之后就只想一心求安稳,靠拍领导马屁和骗取科研经费站稳脚跟; 而私营企业,在30岁后就被迫走“高级开发”、“管理路线”,不然会被指责“没有上进心”。

私营企业真正应该沉淀的项目管理经验和架构设计思路,完全被超长期工作压得喘不过气; 而由于高校计算机教材的严重脱节,又导致培训机构/付费知识学习横行。

综上所述,计算机类的课程,必定要全电子化。针对少量贫困地区,可以采取纸质印刷。

学习曲线贯彻一生

熟悉我的代码风格的人都知道,我写的代码是一种“混合”的艺术,原型来自于生活,解法也来自于对生活的思考。

就像P=NP问题中的TSP问题。我也是想了10年,才得出用时间去丈量N维线段的长度,从而解题。

但在语法静态分析层面,我还需要学习离散数学,抽象代数和线性代数等。

真正入了这行,才直观了解人与人的差别,跟天与地的差别那一样大。

我只不过是等了5年,才等到今年人工智能大爆发,用N型人工智能完成辅助计算。

实际上我的学历也就高中水平,偶尔无聊去面试还被各种鄙视,言语里面无非就是“你一个本科学历都没有的人还想着入行?”

然后我就只能实际行动告诉他们,为什么在但丁写的《神曲》里面,对他人不敬的傲慢是最大的原罪。

Recently, through DNS, I discovered that this website has more overseas traffic than domestic traffic. A while ago, when I went back to high school, I found that very few people were studying liberal arts. Most of them were crowding into science, especially computer science.

Most People Should Not Blindly Study Computer Science

But in my opinion, including myself, most people are actually not very suitable for studying computer science—I just accidentally chose science and was forced to walk the computer science path all the way to the end.

When I first entered the industry, I studied diligently and worked seriously. But in a third-tier city, it actually didn’t mean much; effort always has a ceiling and limit. And the ceiling of a third-tier city is your own parents.

Although to most people, my talent is not ordinary, so countless people envy, resent, and hate me. But in my view, I write very little code, “lacking a big project,” and my ability ceiling is my time, and I have already started to feel the decline of my body.

After 10 years in the industry, I actually spend more time daydreaming than working. And because of the toxic domestic environment, it’s rare to be truly valued. At first, I despised those low-educated colleagues around me who deliberately smoked next to me; but when I came to a big city, I found that I became the one despised.

The domestic environment lacks basic respect for people, always emphasizing various “control theories” — as a minor, you are kidnapped through authority relationships (parents and teachers are always right, you must listen to me); after graduation, you are limited by your job; when you earn enough money, you can go through the “buy a bigger house, get married, have children” full-package process. Buying one house is not enough; you have to buy one for the eldest son and another for the grandson.

However, material resources are limited, and internet jobs are also limited. Every year, there are millions of college entrance exam graduates; if they all covet the “high salary” of this industry and enter this track, then ineffective internal competition will force this industry to become low-paying.

When going to work becomes a performance, even if there is nothing to do, everyone still has to pretend to work at their desks. I was on the edge of a big company’s OB for a few years and have basically completely demystified them.

The Paradox of the Internet Industry

Moreover, the first paradox of the internet industry is “Kill Everything” — the more excellent the practitioners are, the more antisocial they must be.

The foundation of the internet industry is information technology, and information technology represents a new quality of productivity. Its original setting was to “assist humans,” but gradually it turned into “replacing human jobs,” such as open-source projects devouring closed-source software companies, Kubernetes replacing manual operations, and AI destroying grassroots programming jobs (studying computer science at a vocational school is no better than learning car repair).

In China, basically all internet businesses have been fully developed, but there are few internet companies—people generally lack a spirit of interconnection and always think of building “big platforms” behind closed doors to let others join, then collect rent as “landlords.”

But now that AI has emerged, we have a chance to “set things right.” When formulating AI ethics-related laws, I think “people-oriented, machines as auxiliary” must be placed first.

Moreover, computer-related textbooks must start to be rewritten.

The People Who Make Textbooks Do Not Understand Computers at All

In the Python language, indentation represents nested levels. If following the existing approach and printing with paper textbooks, indentation becomes blank space, which actually reduces readability.

Not to mention the domestic textbooks are a mess, and the quality is really hard to describe. Outdated language books are not taken down for a long time, and the authors’ thoughts are backward and inflexible. The content they write is as hard to understand as a sacred text, only making people want to sleep in class.

The biggest problem with these authors is “laziness.” They think that after muddling through for so many years and finally becoming authorities in education, they can decide the content of textbooks, and their thinking becomes lazy, no longer embracing new changes.

On the other hand, language standards themselves are changing. Although the ancient C and C++ languages are still shining, these obscure and difficult languages are obviously not suitable for most beginners.

C# is elegant but lacks an ecosystem; Java claims to be cross-platform but is like a clown in the cloud-native field (Java in docker); Go claims simplicity but is far from as elegant as C#’s linq + lambda in terms of generics and collection operations compared to 10 years ago.

And computers have been developing for more than 30 years, but our only programming language is an “internal fighting language.” We only achieve unprecedented unity in internal consumption of this problem.

Those talents who study computer science, after getting a university teaching position, only want to seek stability, relying on flattering leaders and cheating research funds to secure their positions; and in private enterprises, after 30 years old, they are forced to take the “senior developer” or “management route,” or else they will be accused of “lacking ambition.”

The project management experience and architecture design ideas that private enterprises should really accumulate are completely suffocated by overwork; and due to the serious disconnection of university computer textbooks, training institutions and paid knowledge learning run rampant.

In summary, computer courses must be fully electronic. For a small number of poor areas, paper printing can be used.

The Learning Curve Lasts a Lifetime

Those familiar with my coding style know that my code is a kind of “hybrid” art, with prototypes coming from life and solutions coming from reflections on life.

Just like the TSP problem in the P=NP question. I also thought about it for 10 years before concluding that using time to measure the length of N-dimensional line segments can solve the problem.

But at the syntax static analysis level, I still need to learn discrete mathematics, abstract algebra, and linear algebra.

Only after entering this field do you intuitively understand the difference between people, as big as the difference between heaven and earth.

I just waited 5 years until this year’s AI explosion, using N-type AI to complete assisted computing.

In fact, my education is only at the high school level. Occasionally, when bored, I go for interviews and get despised in various ways, with words like “You don’t even have a bachelor’s degree and still want to enter the industry?”

Then I can only show them through actions why in Dante’s Divine Comedy, arrogance that disrespects others is the greatest original sin.