Post by @nina_kali_nina@tech.lgbt

2026-02-22 12:06 UTC

@gsprs@mastodon.social I'm well aware that the "hidden" code can be read if one cares about it. But the UX is bad, and large changes go unnoticed for someone who only skims over the PR. Which is more or less the only option for a PR that changes 161 file. These PRs are generally very difficult to review for humans. > Also, who cares if master breaks? Every reliability engineer worth their salt.

Replies (2)

@gsprs@mastodon.social 2026-02-22 13:44

@nina_kali_nina@tech.lgbt > These PRs are generally very difficult to review for humans. Is it difficult for humans using AI? I’ve heard it’s easier that way 😁 The anti-AI crowd is more than welcome to put in the work and fork the projects they criticize for using LLMs and maintain their own repo with 100% organic homegrown code, I wouldn’t hold my breath waiting for that though, being outraged over other people’s generous contributions is far more attractive.

Open ##5221390
@lispi314@udongein.xyz 2026-02-24 05:37

@nina_kali_nina@tech.lgbt @gsprs@mastodon.social Honestly I'd be inclined to just reject the PR if it's not very pertinent. Even then I might procrastinate on having the energy to actually review that. I refuse to just merge in code I didn't actually review.

Open ##5221391