You're reading the public-facing archive of the Category Theory Zulip server.
To join the server you need an invite. Anybody can get an invite by contacting Matteo Capucci at name dot surname at gmail dot com.
For all things related to this archive refer to the same person.
Tim Hosgood (Feb 27 2026 at 15:57):José said:
I am not a native English speaker. Indeed, English is my third language, after French, my second language and Spanish, my first language. Therefore, I need to use LLMs with the following prompt: "Write the following text as a native English speaker".
you might find #community: espanol/Spanish and #community: francais/French good places to start a discussion then :-)
Notification Bot (Feb 27 2026 at 15:58):5 messages were moved here from #learning: reading & references > 📚The Stacks Project (Pedagogy) by José.
José (Feb 27 2026 at 16:10):Tim Hosgood said:
you might find #community: espanol/Spanish and #community: francais/French good places to start a discussion then :-)
I appreciate the thought, Tim, and I know you’re trying to be helpful. However, I believe true inclusion means using tools like AI to help non-native speakers participate directly in the main conversation. To me, it feels more productive to use technology to bridge the gap here in the primary channel rather than moving to separate, language-specific spaces.
James Deikun (Feb 28 2026 at 01:16):Chatbot AIs are not really appropriate for translation (or creative writing), both because they are too biased toward not breaking their personas, which covers up a lot of the subtleties of expression a good translation would convey, and also because their interface makes it difficult to gauge how much is actually being understood. They produce fluent text even when their translation is inaccurate, which doesn't give either party any clue about what is being lost in translation.
Kevin Carlson (Feb 28 2026 at 01:26):I would add that they're very excellent for translation of basic statements between reasonably arbitrary pairs of languages, but I agree with those risks of trying to use them to communicate subtle philosophical and mathematical ideas across a language barrier. In that case I think hand-written language, even if non-fluent, is almost always going to be more effective.
David Michael Roberts (Feb 28 2026 at 01:26):Why not use a traditional translation site/service? Google translate/DeepL? They don't confabulate or use a very recognisable style that grates.
José (Feb 28 2026 at 01:56):The goal is not just translation, but writing "as a native English speaker".
Kevin Carlson (Feb 28 2026 at 02:07):But why would you want to simulate a native English speaker when you are not one? That's misleading your interlocutor about yourself in a quite fundamental way.
James Deikun (Feb 28 2026 at 02:12):Chatbots are specifically designed to be unable to express many sentiments, connotations, and forms of speech which a native speaker would naturally use. Therefore it cannot translate your writings into the forms a native speaker would use. Further, as they advance, they are likely to get worse at this, not better, because the failure is designed in. And asking an AI to interpret, rather than translate, while obscuring its presence and precise role in the conversation is fundamentally dishonest.
David Michael Roberts (Feb 28 2026 at 05:18):As we've said to people on MathOverflow: we prefer imperfect English from your own writing and your own ideas, rather than polished AI-bland drivel.
David Michael Roberts (Feb 28 2026 at 05:19):Somehow, people managed to communicate mathematics in broken English (and French, and German, and....) for decades.