On Tue, Jun 2, 2026 at 6:50 PM Go Kudo <[email protected]> wrote:
> > > 2026年6月2日(火) 23:53 Alexandru Pătrănescu <[email protected]>: > >> Hi Go Kudo, >> >> I am writing to you privately regarding your recent emails on the >> internals thread. >> >> I appreciate the use of AI tools to ensure clear and professional >> communication; I often use them myself when working in other languages. >> However, I’ve found that the replies have become quite long, making it >> difficult to stay focused and read them entirely. >> >> If you are using an AI to generate these responses, would you consider >> asking it to be more concise or to provide shorter summaries? >> Or maybe better, try using it primarily for translation and refinement of >> your own notes rather than letting it generate the full bulk of the content. >> >> I apologize if I am mistaken and you aren't using AI for these replies, >> but I wanted to share this feedback in case it helps. >> >> Best regards, >> Alex >> >> > Hi, Pătrănescu. > > Thanks for your advice. > > As you pointed out, I have been using an LLM to compose my replies up > until now. > This was because my English skills are insufficient, and I was afraid I > wouldn’t be able to convey what I wanted to say. > > But, your opinion is true, Text is very long and redundant. That makes it > unreadable. > > Currently text is written and does not use LLM. but my English skills are > too bad. > > However, I understand most people's dislike LLM text, and I understand the > dislike of LLM use for development. > RFC Implementation is all human reviewed & tested, but porting > implementation of our `colopl_cache` extension with LLM. > > If these LLM use cases are not acceptable to a PHP organization or > community, I'm zero-scratch retiring from implementation and discussion. > > I want to hear your opinion. > Just to be clear, my feedback is only about your messages on internals mailing list discussion. And related to that, from my point of view, I have no issue with you using LLMs for writing. I would even say that it clarifies your messages compared to a few months ago, making sure it goes in a well-articulated English language format. But try using LLMs more for translating and correcting your text and ideas and less for expanding and generating large blocks of text. Shorter, more concise text is better. We need to be considerate of the people reading the messages and make it easy for them as well. Keep up the good work, Alex
