@jwildeboer Re: "data hoarders will collect whatever training data"

Wait, are you considering "OS models" made with non-public data? If the data isn't either in public domain or on a permissive license how do you want to legally distribute a derivative work generator with traces of that training data on such a license?

@jwildeboer The training data is not there for recreating the model. The training data is there so you have access and can validate that indeed all the ingested data is either CC0, or some other license and thus derivative work produced with that model which adheres to this other license fulfills the requirements of the training data's license.

You don't have to move both together until you fork it with the intent of distributing it.

[1/n]

@jwildeboer Isn't the training data requirement a basic liability CYA?

Unless the person providing the model takes on full responsibility (fiscal or otherwise) for any copyright and patent infringements of the model from now until eternity (I don't think you can indenture your children and grandchildren like that, can you?).