AI-written code is riddled with bugs

I recently found a program developed by a skilled developer using AI—it’s actually a pretty complex piece of software, but my experience with it was full of issues everywhere.

This program is packed with features: an online store, forum, blog, navigation, documentation—you name it, it’s got it.

But in reality, when you use it, there are lots of logical errors and missing features, so it’s basically full of problems.

If the code had been written manually, these issues would likely not have happened.

Conclusion

When using AI to write code, don’t aim for something big and all-encompassing—go for something small and refined.

Trying to cover everything just means AI will leave you with massive, hard-to-find bugs.

The ideal is to create something small and beautiful, focusing on doing one thing really well.

2 Likes

This isn’t AI’s fault, it’s the product manager’s. Logical inconsistency is a problem with product design—of course, if the AI didn’t promptly fill in the gaps, it also means this AI isn’t up to par.

3 Likes

Haha, if the product manager doesn’t use AI well, you can’t blame the AI.

1 Like

Everything James said :xhj16:

1 Like

Yes.

1 Like

It could also be that the AI is just kind of dumb.

1 Like

An 8B model writes code

AIGC ≠ Human-Written

This is the fault of the person using AI ~.~