The real issue would be the quality of the proposed solution - whatever the tools or system used to produce the proposed code change. It seems that the code has been examined and the quality bar has passed satisfactorily. ( My interest in this is that I was a test analysist for 20 years for mainly corporate software . I did not do code reviews )
Makes sense to me. What's the difference between a maintainer reviewing code coming from strangers and reviewing code coming from an AI? I remember a story where university researchers were proud to have merged backdoors into Linux, it's not like humans are not adversarial.
I don't like that AI is consuming a lot of energy and reinforcing inequalities (and making oligarchs more powerful), and all that. But the fact that a good maintainer accepts to merge code coming from an AI that they have reviewed doesn't surprise me at all.
I am a maintainer myself (of smaller projects of course), and I can say that LLMs today write much better code than the average contribution by a random human... In fact the quality of the contributions I have been receiving has increase, because people use AI. It is still my responsibility to understand the code I merge, but nothing has changed there.
The real issue would be the quality of the proposed solution - whatever the tools or system used to produce the proposed code change. It seems that the code has been examined and the quality bar has passed satisfactorily. ( My interest in this is that I was a test analysist for 20 years for mainly corporate software . I did not do code reviews )
Makes sense to me. What's the difference between a maintainer reviewing code coming from strangers and reviewing code coming from an AI? I remember a story where university researchers were proud to have merged backdoors into Linux, it's not like humans are not adversarial.
I don't like that AI is consuming a lot of energy and reinforcing inequalities (and making oligarchs more powerful), and all that. But the fact that a good maintainer accepts to merge code coming from an AI that they have reviewed doesn't surprise me at all.
I am a maintainer myself (of smaller projects of course), and I can say that LLMs today write much better code than the average contribution by a random human... In fact the quality of the contributions I have been receiving has increase, because people use AI. It is still my responsibility to understand the code I merge, but nothing has changed there.
[dupe] https://news.ycombinator.com/item?id=50090842