x
Hard vs Soft in fields as attitudes towards model collision — LessWrong