Hacker News new | ask | show | jobs
by yorwba 113 days ago
Section 6, "Removing Watermarks," of the paper you cite makes it very clear that detecting LLM-generated text is not solved if the user takes measures to avoid detection.