AI ruined my computer science degree

AI is ruining my computer science degree.

The person who wrote this complaint is Katie, a sophomore girl who has dreamed of studying computer science since she was a child, Katie.

She has no elite school background, only got a B in high school math; she never received any formal professional training before, and the only decent "project experience" she has is just a few slightly crude web-based mini-games.

She originally thought that after stepping into the ivory tower of university, she could finally study computer science properly!

Who could have imagined that——

The course syllabi are generated in batches by Claude, the final assignments are automatically graded by LLM, and even the caring emails sent by her supervisor are very likely to be generated by background automated scripts...

Faced with this absurd scene, she collapsed (just kidding):

Thanks to AI, my university seems to be going through a life-or-death crisis.

Less than 24 hours after the blog was published, it quickly climbed to the top of the hot list on Hacker News homepage.

The Computer Science Department That Has Completely Collapsed

In her learning experience, since the outbreak of this wave of generative AI, the school's teaching policies have never stopped swinging back and forth.

At the very beginning, they fully embraced AI of course:

Grading students' assignments and writing comments, generating course syllabi, sending mass reminder emails to urge students to study... all these tasks are assigned to large language models.

And what was the result?

After less than three months of tossing around, they found all the outputs were full of hallucinations and garbage content, so they withdrew the policy immediately.

The entire teaching management system has always been in an extremely schizophrenic state——

You must not use AI at all!

Oh no, I got it wrong, everyone must use AI!

No no no... it seems that AI still cannot be used...

Then a very surreal thing will happen.

For the same course, the professor forces you to use AI for the assignment of the previous module, saying that it is an important assessment dimension for the course grade.

After hearing that, you gritted your teeth and paid out of your own pocket to subscribe to Claude Plus.

When you get to the next module, you upload the code carefully tuned with Opus 5.5 full of confidence, only to find that——

What the heck?

What do you mean by academic misconduct, and using LLM will lead to point deductions???

But to be honest, if it's just the ever-changing rules, it's not that bad, you can barely cope with it.

What breaks the author most is that she studied hard to get into university and paid such high tuition fees, just hoping to learn systematically from expert professors.

But it turns out, you, a university professor, are also fooling me with Claude???

Yes, the instructor on the podium sternly warns students not to take shortcuts, then turns around and directly throws the requirements to Claude to generate teaching materials with one click.

In the past, assignments were at least carefully considered by professors combined with teaching objectives, and reviewed by the teaching research group. The task volume was moderate and progressive step by step.

Now? The girl suspects that those professors haven't even opened the markdown file.

The whole document is full of nonsense generated by large models. The 3000-word assignment instruction requires submitting 15 pieces of materials.

After doing the math, the workload of assignments has tripled compared to before...

As a result, ordinary students who already have a weak math foundation and transferred from other majors are forced to skim through the knowledge points quickly.

I came here to learn real things steadily, damn it, not to float on the surface of 15 impossible tasks in a 3000-word assignment!

But this is not the most frustrating part.

Many students worked hard for more than ten days, but for some reason, the code just couldn't run no matter what.

One week before the DDL, the teacher suddenly sent a correction email:

There are two mistakes in the syllabus, which have been corrected now. Please resubmit your assignments.

All the Tokens I burned in vain😭

So why on earth are they doing this?

Katie said, it can't all be blamed on the professors. The real culprit is the management.

Do you know what capital loves to hear most? Scalable learning, preferably automated!

In this way, they can squeeze as much profit as possible from tuition fees without paying extra staff costs.

Since large language models can automatically generate exam questions and grade assignments in seconds, why spend a lot of money hiring teaching assistants and increasing the number of lecturers?

Technology has become an excuse for the school to cut teaching expenses infinitely, squeezing lecturers endlessly, and the workload is so large that it cannot be completed only with human attention.

Katie also mentioned that she can hardly see her supervisor at ordinary times.

The only contact she has is that when she doesn't log in to the online learning platform for several days, she will receive a letter from the supervisor——

She opened it full of nervousness, only to find that it wasn't written by the supervisor himself, it was triggered by the automatic alert o(╥﹏╥)o.

The Most Viral Post on Hacker News

After the post was published, it quickly went viral all over the Internet and climbed to the top of Hacker News homepage.

Many computer science practitioners poured into the comment section.

But to my surprise, even though they are all from the computer science department, not all voices are empathetic.

The entire computer technology community, is split into two opposite poles.

One group is senior veterans who have long established their positions in the industry.

From their perspective, all the current chaos is just a short-term pain in the evolution of technological paradigm.

I have been engaged in this profession for more than 20 years, and I can learn new things every day. In the past few years, with the development of artificial intelligence technology, my learning speed has been faster, and I can understand the systems I work on more deeply.

All in all, AI is just a handy tool. As long as you make good use of large language models to conduct Socratic-style guided learning, the current learning resources are far better than those in their era.

Artificial intelligence will not replace education and traditional learning methods.

On the contrary, artificial intelligence will improve your ability to understand and memorize. As a self-taught professional with strong technical thinking, what I have always longed for is to keep learning.

Therefore, many experienced seniors even have a little bit of envy for the new generation:

Becoming a senior engineer does not really depend on how long you have been developing software. I have seen some very outstanding 21-year-old developers whose programming ability can even surpass those who have worked for more than 20 years.

I think the reason why they can perform so well is that they have solid basic knowledge of computer science.

But this is after all a narrative that only belongs to a small number of people.

On the other side, the vast majority of young people who just graduated, only feel suffocated.

A commenter posted the conclusion of a quantitative study:

Researchers from Stanford University found that the employment trend of junior employees they predicted last year still exists and is expanding.

Specifically, the employment level of 22 to 25-year-old employees working in industries most vulnerable to AI impact is 19% lower than that of their peers who rarely face AI shocks.

In response to this, some practitioners who have been involved in recruitment for many years also frankly confirmed this reality:

Indeed, it is more efficient to equip senior engineers with top-level Coding Agents than to spend time training a fresh graduate from scratch.

I really don't know what advice to give to developers who are about to graduate.

Spending four years of youth studying computer science, but in the end can't even find a job……

Then, what on earth is the meaning of a CS degree?

Retrieve the Warmth of Programming

As for this question, I think the answer may be found in Katie's blog post.

At the end of the article, she wrote this passage:

I don't care if computer science and mathematics have reached their end! I still want to learn!

Okay, I do care about my future career development, but that doesn't mean I want to forget everything I have learned, or stop exploring.

There are so many interesting directions worth exploring, some related to AI, some not. But overwhelming headlines are always talking about things like "Software Engineering is Dead" and "You will be eliminated if you don't use AI", which makes people breathless.

Can't a girl just learn statistics purely for fun?

After going through all this, Katie's answer is——

I can keep going, but I won't. It will give me indigestion.

Maybe I will go back to OpenProcessing and make some sillier games.

At least now I really know how to do some programming.

Then she posted a few crude small web pages she wrote on OpenProcessing in her early years:

A simple space shooting game with exciting background music;

A Moogle that will make a "meow" sound when you click it;

And an interactive demo that scares small fish away with the sound from the microphone.

To be honest, they are very rough.

Any one of them is far worse than the effect generated by a single Prompt of Astra or Fable today.

It may not even be as popular as the pelican bicycle demo of a new model...

But the moment you open it, you can feel something that models can never generate——

Vitality.

That's what many programmers looked like when they first fell in love with writing code.

It's not to brush hundreds of LeetCode problems on the job application assembly line, not to report how much Token call cost you saved in the weekly report, let alone to cater to the efficiency indicators randomly put forward by middle managers of big tech companies.

Before all of this was disciplined by utilitarianism, originally just typing a few lines of characters in the terminal with black background and white text, watching the lit pixels on the screen run and flow according to your own will, was enough to make you too excited to sleep all night.

We have personally created super machines that know almost everything.

But it seems that unconsciously, we have locked the most primitive human desire for exploration and sense of immersion out of the door……

Original blog post: https://lordtoots.bearblog.dev/ai-is-destroying-my-cs-degree/

This article is from the WeChat official account QbitAI, author: Focus on cutting-edge technology, published with authorization from 36Kr.