Quantifying the algorithmic improvement from reasoning modelsepoch.ai1 point·famouswaffles··0 commentsOpen articleSaveView on HN