zlacker

Would it even be possible for OpenAI to excise the NYTimes data from their models without running all the training again? Seems like a huge mess, particularly since they'd have to do that each time they lose a lawsuit.