I don't like reading content that has not been generated with care. The use of LLMs is largely orthogonal to that. If a non-native English speaker uses an LLM to craft a response so I can consume it, that's great. As long as there is care, I don't mind the source.