How is that BS? If the technique works, you can grow your sample of imaged brains and viewed images ad infinitum, and then why wouldn't I be able to tell what random thing any random person is looking at?