Results similar to mythos have been duplicated by weaker models.

Think it's more a care of mythos raising widespread awareness that tireless LLMs can be weaponized to dig through code and find that one tiny flaw nobody spotted