"Dr. Arne Babenhauserheide" <[email protected]> writes: > Ihor Radchenko <[email protected]> writes: >>> So if we accept code that the submitting human does not understand, we >>> can reach a state that’s unmaintainable for humans. >> >> I think that this point is the most important. >> You somehow assume that we will be accepting code that humans cannot >> understand. We will not. > > I wouldn’t phrase it as "cannot" but at "do not". > > As long as we expect the contributor to understand the code they submit, > I don’t worry too much.
> I mostly worry that contibutors will stop reading the code and expect > others to review code the contributors never read themselves (or wave it > through "because AI"). +1 > Because by now that’s what everyone I know personally who uses AI has > ended up doing. Even the one who I thought didn’t do that. That's up to some point. Until the users run into major issues that pop up from unreviewed LLM-generated code. I have to review such code from students, so I've seen a number of examples. > Or if people start saying “let AI do a pre-review” -- that just means to > force contributors to read AI output. If I as reviewer don’t want to > read unchecked AI output, I shouldn’t force contributors to read such > either. What about pre-review only for LLM-generated code? (That's what I did for John Wiegley's patch). An alternative could be providing LLM usage guidelines, but that may be too much. -- Ihor Radchenko // yantar92, Org mode maintainer, Learn more about Org mode at <https://orgmode.org/>. Support Org development at <https://liberapay.com/org-mode>, or support my work at <https://liberapay.com/yantar92>
