컴퓨터 프로그래머

MR이 말합니다.

Hi! From which year do you use data from and when was the calculations performed?

Darin Beaudreau (적당한)이 말합니다.

20 years is a reasonable timeline for a large portion of software engineering to be able to be automated, but the problem in replacing programmers isn't AI's ability to reason, but in the customers themselves. To be able to fully replace engineers, a customer needs to be able to accurately describe their requirements, and anyone who works in that field will tell you that customers are idiots.

I think web development will be the first variation of software development that is automated because the majority of all web development is the same and very basic. Things like embedded, robotics, and applications programming will take much longer and require AI to be able to see the whole of an application in a single context to avoid a plateau effect where you get increasingly more errors due to loss of information regarding the program's architecture.

user (적당한)이 말합니다.

I've been a software engineer for over 10 years. This is really only the beginning. The pace of innovation in AI will only increase. We are already using Github Copilot at work and AI gives amazing code reviews. Outsourcing is now much more effective. Less engineers are needed already because of AI. With "Devin", we are seeing a tiny glimpse of what is to come in the next 20 years. I'm going back to school and transitioning out of the industry over the next 2-3 years.

maikelos272@gmail.com (매우 가능성이 높음)이 말합니다.

Although AI is a tool for programmers, the better it gets the more of our job it can perfrom. It is unfathomable to me that AI will not be able to do a programmers job really well within the next 20 years. I reckon it will happen in 10. The only hope I have is that rather than the job market deminishing, its output will grow exponentially and programmers will grow to be supervrisors of AI. Nontheless at some point AI will be able to write apps and programs based purely on Natural language input. Albeit the code might be far from perfect at the beggining it will only get better.

ezkha (불확실한)이 말합니다.

So I've heard about this AI "Devin", which is apparently capable of programming and distributing an entire application from scratch to release; but it is absolutely terrible at managing issues and fixing bugs. If my source is credible, it's 20% efficient compared to 80% with a human.

I think it'll probably improve and replace programming. Though maybe issues and solving bugs, in particular, will take a while to replace.

user (매우 가능성이 높음)이 말합니다.

AI has been on the rise in the past several years, and is likely to continue growing rapidly. Within the past years what has been possible has greatly increased many magnitudes, and several AI tools that can accomplish coding means have been created, and are being greatly invested on and improved. With the release of "Dave" and companies like "Magic.dev" I can foresee this being a greatly human reduced job in the next 2-6 years.

S (매우 가능성이 높음)이 말합니다.

ChatGPT is already proficient. It will at the very least heavily reduce the amount of software engineers required

Johnifer (기회 없음)이 말합니다.

Those who say that AI will completely take over CS don't know CS.

Jon (불확실한)이 말합니다.

I see that most people voting with certainty on this topic don't understand how this job works, or they are just average or bellow average professionals.

GPT really gives you solutions for complex tasks but on minor details it's going to invent some buggy solution ultimately creating more bugs than helping you.

That's why you really need to understand the tool/system that you're working with.

kir (낮은)이 말합니다.

AI has a difficult time with complex use cases, architecture, and requirements from clients.

Rin (매우 가능성이 높음)이 말합니다.

For being beings based on technology, programs, and algorithms, computer programming will be able to replicate itself and take over its own recreation processes.

Ryan (적당한)이 말합니다.

Basic functionality in software is easily automated. Building entirely new systems might be more difficult. If we develop AGI, then any job done at a keyboard is gone, but if that proves more difficult than expected then high level system design and development jobs might last a bit longer.

Bryan Schmidt (적당한)이 말합니다.

So much of what we do as computer programmers can be automated.

My job is to interpret what the product owners want, and convert that into code. Currently, I have transitioned into using AI to help me with most of the code, while I'm still guiding the architecture, code style, and overall codebase.

It won't be long until AI could handle most of if not all of that as well.

John (매우 가능성이 높음)이 말합니다.

Because chatgpt can already make script, sure its a wrong sometimes. but just imagine but it will be like in 20 years.

Auditormadness9 (매우 가능성이 높음)이 말합니다.

It already wrote like 20 scripts for me that completely worked as expected.

Sequenze5517 (낮은)이 말합니다.

Because Computer Programmers made and update AI. Logically, the AI will not update himself.

me (기회 없음)이 말합니다.

coding requires a lot of skill and thinking which cannot be replaced by ai

matin (낮은)이 말합니다.

robots make robots?

Leo (매우 가능성이 높음)이 말합니다.

so clearly they already have stuff like Chat GPT that tells you code,

Anonymous (불확실한)이 말합니다.

As it stands, an AI gets its information from the internet. That means that much of the code it acquires at this point when you ask for sample code or something similar will likely be from GitHub or some similar website. It doesn't have to understand the workings of the code, it just needs to know what to look for and dispense it to you. Sometimes the code may not even work. AI could eventually understand programming and then it'll be a different story, but AI doesn't really KNOW much of what it says, it just give a definition from the huge databases it has access to.