This whole legal debate exists because of ambiguity. Open source licences were written before large-scale training became a thing. Rather than settle this ambi

This whole legal debate exists because of ambiguity. Open source licences were written before large-scale training became a thing. Rather than settle this ambiguity on court, why not remove it altogether? Personally I think we should be adding new clauses to licences that either explicitly approve or prohibit the use of code for training of models.