0/ I don't understand the argument that training LLMs using copyrighted material is not within the scope of fair use

In school, society explicitly trains future artists, journalists, musicians, etc by studying prior work

Cofounder, Managing Partner @ Multicoin Capital

I do understand it, though not agree with them I think it should be discussed, clarified and classified.

those are different subjects also , for instance twitter would like you to pay for the access to their data for training (if they sell it at all)  and won't care about AI creations being plagiats or not.