this post was submitted on 22 Apr 2026
-9 points (26.3% liked)

Opensource

6102 readers
251 users here now

A community for discussion about open source software! Ask questions, share knowledge, share news, or post interesting stuff related to it!

CreditsIcon base by Lorc under CC BY 3.0 with modifications to add a gradient



founded 2 years ago
MODERATORS
all 9 comments
sorted by: hot top controversial new old
[–] amio@lemmy.world 18 points 3 weeks ago (2 children)

"I used to have to check PRs and with LLMs I implicitly trust there's no malicious shit in them"???

Yeah ok bro

[–] coolie4@lemmy.world 16 points 3 weeks ago (1 children)

I mean... I think they're right though. LLMs aren't intentionally malicious. They're just incompetent.

[–] onlinepersona@programming.dev 7 points 3 weeks ago

End result is the same.

[–] CameronDev@programming.dev 0 points 3 weeks ago

While I still need to review LLM-generated code, I generally don't have to worry about it being malicious the way an unknown contributor's code could be.

[–] AceSLS@ani.social 14 points 3 weeks ago

Prefering a LLM's code over a Human's is just a wild statement lmao

[–] remon@ani.social 6 points 3 weeks ago

That's a lot of text for saying "I suck at code reviews".

[–] aev_software@programming.dev 6 points 3 weeks ago

How much time does it take to reject a pull request because you don't understand it?

We see a project like curl outright banning AI-made / assisted PRs due to the overwhelming burden those put on the human reviewers. It's a DDoS aimed at people, either to shut them down or to force them into using yet another LLM to automate reviews and merges.

The danger is real. Overwhelming the people who have maintained the backbones of our digital infrastructure will result in a total collapse of that infrastructure. How can ad brokers continue to profit if there's no way to get the ad displayed to prospective consumers? How can totalitarian regimes sway public opinion in a foreign nation if they can't share their memes and propaganda?

[–] aev_software@programming.dev 1 points 3 weeks ago

How much time does it take to reject a pull request because you don't understand it?

We see a project like curl outright banning AI-made / assisted PRs due to the overwhelming burden those put on the human reviewers. It's a DDoS aimed at people, either to shut them down or to force them into using yet another LLM to automate reviews and merges.

The danger is real. Overwhelming the people who have maintained the backbones of our digital infrastructure will result in a total collapse of that infrastructure. How can ad brokers continue to profit if there's no way to get the ad displayed to prospective consumers? How can totalitarian regimes sway public opinion in a foreign nation if they can't share their memes and propaganda?