Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Using LLMs to write code is a waste. If its rote code or boilerplate it sucks at writing it consistently.

If its novel code, it's not in its training set.

What LLMs don't suck as much is writing semi readable docs and commit messages.



We've got an LLM analyzing merge requests. While most of what it writes is too wordy for my liking and I'd prefer if people making MRs would actually write good descriptions and commit messages instead of saying "we use squash merges", one thing I do like it is that it suggests a better title. The title of an MR is used in the squash merge commit, and that commit is used in the changelog which is what users of our library rely on to consider what they need to check before upgrading.

On the one side, people should write good MR titles for this reason. On the other, sometimes you just don't know, and people's brains are already overflowing with all of the tasks they need to do in order to do their job properly.

But also, finally, it's just a suggestion; we don't have AIs writing our changelog yet, except ironically to make them worse / cringe.


Your understanding of LLMs is wrong. Please read https://www.anthropic.com/news/tracing-thoughts-language-mod...


Respectfully his understanding of LLMs have nothing to do with the paper you linked.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: