Hacker News new | past | comments | ask | show | jobs | submit login

Has anyone successfully repro'd the plaintiff's claims of unlawful use? I just made several attempts using GPT-4 and was unsuccessful.

Perhaps OAI have already taken action to disable ChatGPT's ability to quote in full without attribution, if it was ever possible in the first place(?) Plaintiff doesn't disclose exactly which prompts they used to give rise to their infringement claims, which seems suspect.

If the case goes to trial, and OAI has taken steps to mitigate the harm - i.e., it no longer reproduces NYT content word-for-word - does that significantly reduce the damages they could seek?




I think you need to use the API to reproduce. I heard the chat prompt stops gpt4 from some of this behavior.




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: