Is this a reaction to Apple paper showing that reasoning models don’t really reason?