They already did, obviously. It's not a secret nor aaa is opposed to it, they literally state that they are an archive for llm training as well as a library.

True. Let’s create a data model capable of generating bit-accurate files based on the originals. Then it is not just a copy. It is actually new data, produced by AI. Check mate.

Also lol for trying to delete things from the torrent network.

That is how meta trained llama