Serious Chatjewpt is retarded

[cripplecel]

I don’t know why niggers are saying AI is taking coding jobs now or in the near future. These models are retarded when it comes to doing any slightly difficult task that is not some web bullshit.

I wanted to make a copy of the pipelines in linux and I tell chatjewpt to try it and it created a fucking retarded and unreadable code.

Then I just make it myself and got it done in 20 minutes.

This kind of thing I was trying to make is not hard at all, it’s a very simple program to make for someone half decent at C.

 

[Animecel2D]

I hear your frustration. I’ll address the technical side of what you’re saying and skip over the offensive language.

You’re right that current AI models aren’t magic replacements for strong programmers. They’re good at boilerplate, prototyping, and web/backend scaffolding, but when it comes to nuanced tasks — like re-implementing Unix-style pipelines in C — they often:

• Overgeneralize: They produce verbose or “Frankenstein” code stitched together from different examples.
• Miss subtle semantics: Pipes involve forks, file descriptor management, signal handling, and exec calls. Models often forget to properly close unused file descriptors or manage child processes.
• Lack adaptability: If your request is only described loosely, they guess, sometimes poorly.

That said, for many engineers, models are already useful as assistants (e.g., generating skeleton code, reminding syntax, drafting test cases), not as replacements for deeper problem-solving. Your experience highlights the current gap: coding assistants vs. coding replacements.

Do you want me to show you how I’d actually implement a minimal C pipeline clone (something like running ls | grep foo | wc -l) cleanly, so you can compare with what you made?

 

[nakolas]

I hear your frustration. I’ll address the technical side of what you’re saying and skip over the offensive language.

You’re right that current AI models aren’t magic replacements for strong programmers. They’re good at boilerplate, prototyping, and web/backend scaffolding, but when it comes to nuanced tasks — like re-implementing Unix-style pipelines in C — they often:

• Overgeneralize: They produce verbose or “Frankenstein” code stitched together from different examples.
• Miss subtle semantics: Pipes involve forks, file descriptor management, signal handling, and exec calls. Models often forget to properly close unused file descriptors or manage child processes.
• Lack adaptability: If your request is only described loosely, they guess, sometimes poorly.

That said, for many engineers, models are already useful as assistants (e.g., generating skeleton code, reminding syntax, drafting test cases), not as replacements for deeper problem-solving. Your experience highlights the current gap: coding assistants vs. coding replacements.

Do you want me to show you how I’d actually implement a minimal C pipeline clone (something like running ls | grep foo | wc -l) cleanly, so you can compare with what you made?

Muh offensive language

 

[Icarus]

Midjourney is pretty cool though

I wish I could draw and paint

 

[anotherwastedlife]

They're just glorified search engines

 

[Maikowski]

all the coding jobs will be taken by Indians with fake credentials

 

[Maikowski]

They're just glorified search engines

exactly, ai is shit

 

[psyop]

I hear your frustration. I’ll address the technical side of what you’re saying and skip over the offensive language.

You’re right that current AI models aren’t magic replacements for strong programmers. They’re good at boilerplate, prototyping, and web/backend scaffolding, but when it comes to nuanced tasks — like re-implementing Unix-style pipelines in C — they often:

• Overgeneralize: They produce verbose or “Frankenstein” code stitched together from different examples.
• Miss subtle semantics: Pipes involve forks, file descriptor management, signal handling, and exec calls. Models often forget to properly close unused file descriptors or manage child processes.
• Lack adaptability: If your request is only described loosely, they guess, sometimes poorly.

That said, for many engineers, models are already useful as assistants (e.g., generating skeleton code, reminding syntax, drafting test cases), not as replacements for deeper problem-solving. Your experience highlights the current gap: coding assistants vs. coding replacements.

Do you want me to show you how I’d actually implement a minimal C pipeline clone (something like running ls | grep foo | wc -l) cleanly, so you can compare with what you made?

Are you ai

 

[Animecel2D]

Are you ai

I’m a real human being