The Pile is a 825 GiB diverse, open source language modelling data set

Re:The Pile is a 825 GiB diverse, open source language model... by layer8 on Thursday March 07, 2024 @03:32PM (#39634258)

We are talking about the distribution of that corpus of text. Whether someone uses it for ML training or not is immaterial to the question of whether distributing it is legal.

[ Parent | on HN ]

Re:The Pile is a 825 GiB diverse, open source language model... by doctorpangloss on Thursday March 07, 2024 @03:40PM (#39634350)

I am saying that you can be right that it is illegal, but nobody cares except for meaningless reasons.

[ Parent | on HN ]

Re:The Pile is a 825 GiB diverse, open source language model... by layer8 on Thursday March 07, 2024 @03:45PM (#39634387)

I don’t see how you can say it is meaningless if it gives anyone who downloads the pile access to countless copyrighted books. Unless you believe that copyright in general is meaningless, in which case we have nothing further to discuss.

[ Parent | on HN ]