News for hackers. Stuff that matters.
Re:OpenAI's justification for why training data is fair use,... by est31 on Thursday October 05, 2023 @10:40PM (#37786639)
It wouldn't just be expensive, it would basically be impossible to figure out the content owners of random works on the internet at the scale needed for the LLMs to output the high quality content that they do output. Even if they strike deals with the large content owners, they need way way more content than that.
Re:OpenAI's justification for why training data is fair use,... by ang_cire on Friday October 06, 2023 @10:53PM (#37798623)
Too bad. No one has a right to mass data collection.