I feel like the articles on this have been very negative ... but aren't the Anthropic promises on safety following this change still considerably stronger than those made by the competing AI labs?