no it empathizes the importance of training smaller models for longer, like the Mistral "overtrained" models