Article: In the age of AI, computer science is no longer the safe major.

[Anonymous]

Anonymous wrote:Imagine other majors

+100

[Anonymous]

Anonymous wrote:Has anyone been able to read the Atlantic article "So Much for ‘Learn to Code’
In the age of AI, computer science is no longer the safe major"? I'm interested in hearing some thoughts as I look for a place to read it without subscription.

Yes. It’s available on crest.net

[Anonymous]

Anonymous wrote:Has anyone been able to read the Atlantic article "So Much for ‘Learn to Code’
In the age of AI, computer science is no longer the safe major"? I'm interested in hearing some thoughts as I look for a place to read it without subscription.

Articles like this are written by people who have never tried to get ChatGPT to write anything worthwhile. It's all hype. It rarely gets even the simplest things right. It's more of a Grammarly for programming and could never replace a person at least not anytime soon.

[Anonymous]

Anonymous wrote:

Anonymous wrote:Has anyone been able to read the Atlantic article "So Much for ‘Learn to Code’
In the age of AI, computer science is no longer the safe major"? I'm interested in hearing some thoughts as I look for a place to read it without subscription.

Articles like this are written by people who have never tried to get ChatGPT to write anything worthwhile. It's all hype. It rarely gets even the simplest things right. It's more of a Grammarly for programming and could never replace a person at least not anytime soon.

I use OpenAI tools, and others (CoPilot, etc.) daily.

These tools won't replace skilled developers anytime soon, but they *do* get the "simplest things right". They can write short scripts in a variety of languages, complete functions, configuration files, and other short tasks. If I don't know how something works, I can get a working example in ChatGPT in seconds, thus improving my productivity.

More productive programmers won't reduce jobs - it will increase them. After all, if a given vocation suddenly becomes 2 to 3x more productive, why would industry reduce resources invested?

[Anonymous]

Anonymous wrote:"Coding" is only a small part of programming. I do it for a living. Learn to code was always a joke, because teaching someone to write doesn't make them a novelist.

This. When I hire developers, I’m looking for someone who cares deeply about solving problems with technology and also understands how it minimize the impact of human error. It’s more of a mindset and temperament than a specific set of skills. I work with so many junior people who come into the field with just academic study, and they can’t, for example, fix an OS issue because they never worked on anything outside of their classes. This means that the first time an issue comes up, they are at a total loss and waiting for someone to fix it for them. I recently had a supposedly senior test automation engineer tell me that she was blocked on a task because she “couldn’t open PDFs on her machine.” Those people shouldn’t be in tech.

[Anonymous]

Anonymous wrote:"Coding" is only a small part of programming. I do it for a living. Learn to code was always a joke, because teaching someone to write doesn't make them a novelist.

This. And 4 years of high school Arabic won't prepare you to negotiate peace talks in the middle east.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Has anyone been able to read the Atlantic article "So Much for ‘Learn to Code’
In the age of AI, computer science is no longer the safe major"? I'm interested in hearing some thoughts as I look for a place to read it without subscription.

Articles like this are written by people who have never tried to get ChatGPT to write anything worthwhile. It's all hype. It rarely gets even the simplest things right. It's more of a Grammarly for programming and could never replace a person at least not anytime soon.

I use OpenAI tools, and others (CoPilot, etc.) daily.

These tools won't replace skilled developers anytime soon, but they *do* get the "simplest things right". They can write short scripts in a variety of languages, complete functions, configuration files, and other short tasks. If I don't know how something works, I can get a working example in ChatGPT in seconds, thus improving my productivity.

More productive programmers won't reduce jobs - it will increase them. After all, if a given vocation suddenly becomes 2 to 3x more productive, why would industry reduce resources invested?

Yes, it's very helpful for writing comments like increment i by 1 or coming up with clear variable names provided it's used to review code.

[Anonymous]

Anonymous wrote:If you were getting a university CS degree at a good school fo “learn to code” you were doing it wrong. Actually colleges are doing it wrong because it is hard to learn to code in college if you are not a cs major. But cs is about algorithms and a logical way of thinking that ai will not replace. If that isn’t of interest happily you can learn to code with chat gpt now. At least in theory.

In fact, most CS programs assume you already know how to code in several languages (and if you don't, you need to learn it on your own while doing work in a class using that code), and the really tough classes use 'made up' languages to test concepts, so you can't cheat with known code (the way linguistics classes sometime use made up human languages to test constructs).

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Has anyone been able to read the Atlantic article "So Much for ‘Learn to Code’
In the age of AI, computer science is no longer the safe major"? I'm interested in hearing some thoughts as I look for a place to read it without subscription.

From the article: "Rather, the turmoil presented by AI could signal that exactly what students decide to major in is less important than an ability to think conceptually about the various problems that technology could help us solve."

Which has always been true, and is why the "learn to code / humanities are pointless" rhetoric is so annoying.

The fact is that humanities majors have always learned less in college than STEM majors. Survey after survey points this out.

Business majors learn the least.

Mine is a humanities major with a heavy load of STEM courses, just focused on interesting courses, didn't bother to double major or collect minors. Probably learned way more than most STEM majors with standard load.

And at the end of the day, those STEM majors will have jobs and your kid won’t. But at least there’s the satisfaction of “probably learned way more”, lmao.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Has anyone been able to read the Atlantic article "So Much for ‘Learn to Code’
In the age of AI, computer science is no longer the safe major"? I'm interested in hearing some thoughts as I look for a place to read it without subscription.

From the article: "Rather, the turmoil presented by AI could signal that exactly what students decide to major in is less important than an ability to think conceptually about the various problems that technology could help us solve."

Which has always been true, and is why the "learn to code / humanities are pointless" rhetoric is so annoying.

The fact is that humanities majors have always learned less in college than STEM majors. Survey after survey points this out.

Business majors learn the least.

But who writes the surveys? Social science majors.

[Anonymous]

Anonymous wrote:

Anonymous wrote:If you were getting a university CS degree at a good school fo “learn to code” you were doing it wrong. Actually colleges are doing it wrong because it is hard to learn to code in college if you are not a cs major. But cs is about algorithms and a logical way of thinking that ai will not replace. If that isn’t of interest happily you can learn to code with chat gpt now. At least in theory.

In fact, most CS programs assume you already know how to code in several languages (and if you don't, you need to learn it on your own while doing work in a class using that code), and the really tough classes use 'made up' languages to test concepts, so you can't cheat with known code (the way linguistics classes sometime use made up human languages to test constructs).

In fact, most CS programs assume that you would pick up the language pretty well and pretty quickly rather than that you already know.

For example, my kid's first 2-3 CS fundamental classes were with Java and a made up educational language.
In AI class they use Python, Computer vision class they use C++, in game programming class they use C#.
They quickly go over the language characteristics, and expect you to pick up the languages along the way.
It'll be certainly great if you already know, but not really required.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Has anyone been able to read the Atlantic article "So Much for ‘Learn to Code’
In the age of AI, computer science is no longer the safe major"? I'm interested in hearing some thoughts as I look for a place to read it without subscription.

Articles like this are written by people who have never tried to get ChatGPT to write anything worthwhile. It's all hype. It rarely gets even the simplest things right. It's more of a Grammarly for programming and could never replace a person at least not anytime soon.

I use OpenAI tools, and others (CoPilot, etc.) daily.

These tools won't replace skilled developers anytime soon, but they *do* get the "simplest things right". They can write short scripts in a variety of languages, complete functions, configuration files, and other short tasks. If I don't know how something works, I can get a working example in ChatGPT in seconds, thus improving my productivity.

More productive programmers won't reduce jobs - it will increase them. After all, if a given vocation suddenly becomes 2 to 3x more productive, why would industry reduce resources invested?

Yes, it's saving them time from having to find the right stack overflow page.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:If you were getting a university CS degree at a good school fo “learn to code” you were doing it wrong. Actually colleges are doing it wrong because it is hard to learn to code in college if you are not a cs major. But cs is about algorithms and a logical way of thinking that ai will not replace. If that isn’t of interest happily you can learn to code with chat gpt now. At least in theory.

In fact, most CS programs assume you already know how to code in several languages (and if you don't, you need to learn it on your own while doing work in a class using that code), and the really tough classes use 'made up' languages to test concepts, so you can't cheat with known code (the way linguistics classes sometime use made up human languages to test constructs).

In fact, most CS programs assume that you would pick up the language pretty well and pretty quickly rather than that you already know.

For example, my kid's first 2-3 CS fundamental classes were with Java and a made up educational language.
In AI class they use Python, Computer vision class they use C++, in game programming class they use C#.
They quickly go over the language characteristics, and expect you to pick up the languages along the way.
It'll be certainly great if you already know, but not really required.

I should add that in any case like everyone said, knowing how to code in the languages is a small part.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Has anyone been able to read the Atlantic article "So Much for ‘Learn to Code’
In the age of AI, computer science is no longer the safe major"? I'm interested in hearing some thoughts as I look for a place to read it without subscription.

Articles like this are written by people who have never tried to get ChatGPT to write anything worthwhile. It's all hype. It rarely gets even the simplest things right. It's more of a Grammarly for programming and could never replace a person at least not anytime soon.

I use OpenAI tools, and others (CoPilot, etc.) daily.

These tools won't replace skilled developers anytime soon, but they *do* get the "simplest things right". They can write short scripts in a variety of languages, complete functions, configuration files, and other short tasks. If I don't know how something works, I can get a working example in ChatGPT in seconds, thus improving my productivity.

More productive programmers won't reduce jobs - it will increase them. After all, if a given vocation suddenly becomes 2 to 3x more productive, why would industry reduce resources invested?

I guess you don’t know much about supply and demand and how for profit businesses work.

[Anonymous]

I guess you don’t know much about supply and demand and how for profit businesses work.

You don't get it.

Software development has become far more efficient through the years. In the 70s and 80s, we employed assembly-language programmers. When higher level languages and tools arrived, we didn't simply lay off developers - we just did more.