@mcc @pinkdrunkenelephants I agree that ML models trained on a set of data that contains my code would be a derivative work.

Personally I want my open source work to be used as a public good. A derivative work that is proprietary and used for corporate profit is not for the public benefit, and I find that distasteful.

I want to say that I'd feel less spurned if the ML models were open source, but I don't really know if that's true. Generative models are easily weaponized against the public good as well.