habryka comments on AI Safety Newsletter #41: The Next Generation of Compute Scale Plus, Ranking Models by Susceptibility to Jailbreaking, and Machine Ethics