> Of course it doesn't. That's why the major AI industries are so up in arms about it: because they have to document and describe their foundational models, and

> Of course it doesn't. That's why the major AI industries are so up in arms about it: because they have to document and describe their foundational models, and make sure their data is sound: something that they definitely don't want to do. This is now true for any 1 or 2 person startup. Making it completely impossible to succeed without huge amounts of capital. Congratulations on completely stifling innovation and making life a lot worse for everyone. > Independent devs and open source will be fine: they already (mostly) do that. They are now running the legal risk of the EU commission or some national bureaucrat thinking their models training data are not representative enough or documentation insufficient. How can these things even begin to become judicial questions?