Re:The Pile: a dataset for language modeling [pdf]
by js8 on Wednesday July 12, 2023 @12:21AM
(#36690445)
> If an individual trains a model on their own data to embody their own skills and behaviour, so that they can then sell/rent that model out to work on their behalf
No, because we already do not treat all work as copyrightable. A plumber doesn't get copyright on his piping job. It has to be original enough. So while your own skill might be original enough to warrant copyright, distilling it into a model might not.
Re:The Pile: a dataset for language modeling [pdf]
by tensor on Wednesday July 12, 2023 @10:59AM
(#36695533)
An artists work is copyrightable, a writers work is copyrightable, an a personal model could reproduce those and also produce new works in the same style. Also, data can be intellectual property without being copyrightable.
Re:The Pile: a dataset for language modeling [pdf]
by js8 on Wednesday July 12, 2023 @10:42PM
(#36704440)
> an a personal model could reproduce those and also produce new works in the same style
Yes. So it's like creating a machine that can create art.
Perhaps it shouldn't be copyrightable, but patentable. I think I would be OK with ML models (weights) being patentable rather than copyrightable.
I think it would be a shame to try to build legislation around the notion of the sass melting pot application of machine learning and in the process destroy all sorts of other use cases.