Running 129 Natural Language Processing before LLMs 📝 129 From n-gram counts to attention, learning by doing