Do they though? In lit. 99.999% of works are derivative. Some similar % is for movies. So existing corpus of training data is very likely enough for entertaimen

Do they though? In lit. 99.999% of works are derivative. Some similar % is for movies. So existing corpus of training data is very likely enough for entertaiment purposes. With continued advancemant of models and hardware you basically will get tailored infinite entertaiment content thats optimal just for you. Scientist will continue to advance science possibly with accelerated pace since mudane tasks can be handled by AI. A significant % of resources dedicated to entertaigment can be realocated to more practical pursuits.