Running 132 Natural Language Processing before LLMs ๐ 132 From n-gram counts to attention, learning by doing