Running 136 Natural Language Processing before LLMs ๐ 136 From n-gram counts to attention, learning by doing