It’s almost like, if you train a robot to write text based on books written in the 1950s, then train a robot to detect text written by that robot (who is writing text as if it were a book from the 1950s), then naturally that robot will think a book from the 1950s is written by the first robot.
It’s almost like, if you train a robot to write text based on books written in the 1950s, then train a robot to detect text written by that robot (who is writing text as if it were a book from the 1950s), then naturally that robot will think a book from the 1950s is written by the first robot.