i disagree on this, people comment with stuff anyone could google all the time. its nice to see info without doing it myself, and esp when often i never would. however gpt-whatever is a bad choice for this bc its a very likely scenario to make up bs

> people comment with stuff anyone could google all the time

But presumably with some knowledge or review of it. Commenting with a link to the first Google result comes across about as well as quoting an LLM output.