Hacker News new | ask | show | jobs
OpenAI's flagship AI model has gotten more trustworthy but easier to trick (theverge.com)
13 points by codergautam 978 days ago