News for hackers. Stuff that matters.
Re:The New York Times Launches a Strong Case Against Microso... by js8 on Sunday January 07, 2024 @10:08AM (#38901785)
But if it produces the whole article verbatim, it should be relatively straightforward to match its output to the training set and give attribution, no?
Re:The New York Times Launches a Strong Case Against Microso... by paxys on Sunday January 07, 2024 @10:10AM (#38901810)
If you copy every NYT article and publish them on your own site, writing "this is from the NYT" under it doesn't make it legal.
Re:The New York Times Launches a Strong Case Against Microso... by js8 on Sunday January 07, 2024 @10:25AM (#38901949)
Doesn't make legal what? Copying or publishing the articles? The model output could as easily be modified to output link to the article instead of quoting it verbatim.
Re:The New York Times Launches a Strong Case Against Microso... by paxys on Sunday January 07, 2024 @02:16PM (#38904213)
Okay and if ChatGPT did that then NYT wouldn't have a case. Except that's now how it works.