On social media, Meta’s chief executive appeared to jab at Anthropic by saying that leading A.I. labs should be focused on safety rather than improving their own technology.