Letters to the editor: Study exposes vulnerability of large language models to social media

A study found that large language models exposed to posts from X became less logical, more narcissistic, and harder to fix ...