[IanKerr]

It's been pretty incredible watching these companies siphon up everything under the sun under the guise of "training data" with impunity. These same companies will then turn around and sic their AIs on places like Youtube and send out copyright strikes via a completely automated system with loads of false-positives.

How is it acceptable to allow these companies to steal all of this copyrighted data and then turn around and use it to enforce their copyrights in the most heavy-handed manner? The irony is unbelievable.