Egregoros

@gabriel there's no serious effort to my knowledge to create an open source training corpus.

@sun @gabriel The only people who actually want that or the FSF and FSF supporters, aka none of the proprietary corporations that are currently abusing billions of people morally and technically via the fakeness of opensource.

@sun @gabriel
I'm gonna add that the FSF goal for machine learning should have been the defacto method.
The challenge is that for the moral/ethical path for machine learning you can only use copyleft compatible licensed content to train the models.