This would be better if it fed the parent comment into ChatGPT prefixed with “create a plausible but factually incorrect aggressive response to <comment>”