Many CS concepts taught in unis today are outdated

[Jason Voorhees]

In my previous thread I mentioned about slaving away on DSA and I got asked some questions about why it is is irrelevant so. I'll mention a few more concepts and dinosaurs that are still around and being taught in unis that are more or less worthless.

Like I mentioned DSA doesn't matter unless you are in big tech. Startups, SaaS, and internal tools don't care if your loop is O(n²) .Nobody cared if it was a time complexity O(n²) because we weren't handling millions of requests per second. The priority was shipping features, not optimizing microseconds and modern systems are so ridiculously powerful it doesn't matter anymore but DSA still is one of the most important concepts to learn for interviews and gets asked the most even though the real world uses are very limited.


Most modern bottlenecks aren't algorithmic anymore they're architectural, network, or database related. Yet universities still force feed theory from the 1970s as if everyone's building compilers or operating systems


The sad part is how this the case for many other concepts has become. Disk scheduling, CPU scheduling, compiler Data compression and Huffman coding, design and automata theory all irrelevant with no direct application in modern development. No application and no body bothers with them These are things handled by libraries, kernels, or hardware but students graduate knowing paging algorithms not how to design APIs, write backend systems, deploy on cloud or handle real world debugging. Stuff that actually matters and then niggas wonder why they don't have a job after a degree. We live in a funny time ngl.

 

[59H390]

as a business major

i understand none of this lol

 

[Jason Voorhees]

@User28823 @SoNotFunny @gooner23 @imontheloose

 

[Jason Voorhees]

as a business major

i understand none of this lol

Don't worry many CS grads dont understand it either.

 

[Jason Voorhees]

@menas

 

[Jason Voorhees]

@wishIwasSalludon @5'7" 3/4s

 

[Jason Voorhees]

Im sure there are concepts like this in electrical engineering too @imontheloose

 

[Mitakaa1341]

DSA acts as an IQ filter in universities, weeding out students who can’t handle CS, and in interviews, it's the same.

Learning all of the stuff you mentioned from the bottom up also apparently helps on the job when weird edge cases or hidden issues come up that aren’t obvious from just the top-level view. That's what I was told in a discord server by some nerd who works as a senior engineer, but I'm not personally sure. It could just be him trying to come of as pompous and gate-keepy, idk.

 

[SoNotFunny]

Agreed. Most of the things we are taught are useless. We are better off learning how to effectively use and implement AI tools. Traditional CS is a dead field and it’s a shame that many people are realizing this too late. In fact, you’re better off sticking with IT, mechanical engineering, and anything else hardware related.

 

[wishIwasSalludon]

Just pay 10s of thousands of dollars for outdated lessons theory

 

[Jason Voorhees]

DSA acts as an IQ filter in universities, weeding out students who can’t handle CS, and in interviews, it's the same.

Learning all of the stuff you mentioned from the bottom up also apparently helps on the job when weird edge cases or hidden issues come up that aren’t obvious from just the top-level view. That's what I was told in a discord server by some nerd who works as a senior engineer, but I'm not personally sure. It could just be him trying to come of as pompous and gate-keepy, idk.

DSA isn't some IQ test. Solve enough problems and you'll start noticing the same 20-30 underlying ideas repeating with minor variations. Even those leetcode hard questions are just combine multiple patterns in weird ways or throw in unnecessary fluff. Do enough questions and anyone can get good at it.
I think the main reason it is asked is because DSA is primarily an interview filter it measures your problem solving under pressure.

As for the "bottom-up understanding helps in edge cases" argument I've personally never encountered this.In production I've never had to manually implement trees or heaps. Libraries, frameworks handle that.
Most bugs and "weird edge cases” actually come from system quirks, race conditions, API mismatches, or scheme inconsistencies, not from a lack of understanding of AVL rotations. Bottom up knowledge can help sometimes, but in 99% of real world jobs what you are actually doing is gluing APls together debugging integration issues, and handling system design problems not rriting core CS algorithms.

 

[User28823]

At the end of the day college is just a stamp that says this person can stick to something for four years and knows how to learn, implying they can learn what they need to know on the job

 

[eBoy_]

whats the point of doing CS at this point. There are 100s of international workers that will do it way better, plus AI's getting better and better.

 

[Sprinkles]

In my previous thread I mentioned about slaving away on DSA and I got asked some questions about why it is is irrelevant so. I'll mention a few more concepts and dinosaurs that are still around and being taught in unis that are more or less worthless.

Like I mentioned DSA doesn't matter unless you are in big tech. Startups, SaaS, and internal tools don't care if your loop is O(n²) .Nobody cared if it was a time complexity O(n²) because we weren't handling millions of requests per second. The priority was shipping features, not optimizing microseconds and modern systems are so ridiculously powerful it doesn't matter anymore but DSA still is one of the most important concepts to learn for interviews and gets asked the most even though the real world uses are very limited.


Most modern bottlenecks aren't algorithmic anymore they're architectural, network, or database related. Yet universities still force feed theory from the 1970s as if everyone's building compilers or operating systems


The sad part is how this the case for many other concepts has become. Disk scheduling, CPU scheduling, compiler Data compression and Huffman coding, design and automata theory all irrelevant with no direct application in modern development. No application and no body bothers with them These are things handled by libraries, kernels, or hardware but students graduate knowing paging algorithms not how to design APIs, write backend systems, deploy on cloud or handle real world debugging. Stuff that actually matters and then niggas wonder why they don't have a job after a degree. We live in a funny time ngl.

tech is always evolving

 

[Jason Voorhees]

whats the point of doing CS at this point. There are 100s of international workers that will do it way better,

Most CS grads are unemployable tho.

plus AI's getting better and better.

Reas this

Post in thread 'AI videos now feel 200% more real than what we had just 2 months ago' https://looksmax.org/threads/ai-vid...e-had-just-2-months-ago.1651551/post-23512061

 

[Mitakaa1341]

DSA isn't some IQ test. Solve enough problems and you'll start noticing the same 20-30 underlying ideas repeating with minor variations. Even those leetcode hard questions are just combine multiple patterns in weird ways or throw in unnecessary fluff. Do enough questions and anyone can get good at it.
I think the main reason it is asked is because DSA is primarily an interview filter it measures your problem solving under pressure.

As for the "bottom-up understanding helps in edge cases" argument I've personally never encountered this.In production I've never had to manually implement trees or heaps. Libraries, frameworks handle that.
Most bugs and "weird edge cases” actually come from system quirks, race conditions, API mismatches, or scheme inconsistencies, not from a lack of understanding of AVL rotations. Bottom up knowledge can help sometimes, but in 99% of real world jobs what you are actually doing is gluing APls together debugging integration issues, and handling system design problems not rriting core CS algorithms.

There’s a visible correlation between IQ and performance in competitive programming and LeetCode-style problems. Employers are aware of this, which is why they include such questions in their interviews. Someone with only average or slightly above-average intelligence will generally struggle with LeetCode Hard problems. They might be able to solve a few with enough grinding, but the chances of being asked a familiar problem subset during an interview are slim. On the hand, someone with a higher IQ tends to be more verstaile at doing a wider range of LeetCode-style questions.

 

[5'7" 3/4s]

makes sense, combine outdated practices with ai dependent students, this job market definitely isn’t as cooked as most people think if this is the competition