Stuart_Armstrong comments on [LINK] Normalcy bias