References: Asking well, the anatomy of a good prompt
Source material
Section titled “Source material”This lesson is an original adaptation. Its structure, classroom examples, and quoted lines come from the fourth class of a Harvard Kennedy School course; the prose, framing, and examples for Clawdemy readers are our own. Clawdemy is independent of Harvard, which has not reviewed or endorsed this track.
- The Science and Implications of Generative AI (HKS DPI-681M), Harvard Kennedy School, Spring 2024. Faculty: Sharad Goel, Dan Levy, and Teddy Svoronos. This lesson adapts Class 4: Prompt engineering from the Spring 2024 course site, whose content is licensed under Creative Commons Attribution 4.0. The TIC definitions, the Google-search comparison, the “giving an AI personality” line, the Boston Globe comment-classification story, and the students’ negotiation tips all come from this class’s lecture videos.
- Official course lecture playlist on YouTube, Harvard Kennedy School. The full lectures, free to watch. We mean it when we encourage you to take the original course alongside this track: good teachers deserve more students.
- Chain-of-Thought Prompting Elicits Reasoning in Large Language Models, Jason Wei and colleagues, 2022 (arXiv:2201.11903, NeurIPS 2022). The 2022 research paper the lesson refers to. A precision note: this paper studies prompts that include worked examples containing reasoning steps (few-shot exemplars). The bare walk-through-your-steps instruction, with no examples, is the zero-shot variant studied separately by Kojima and colleagues, 2022 (arXiv:2205.11916).
- FunSearch: Making new discoveries in mathematical sciences using Large Language Models, Google DeepMind blog, December 2023. The lesson’s variability-as-fuel example. The peer-reviewed paper is Mathematical discoveries from program search with large language models, Romera-Paredes and colleagues, Nature, published online December 2023 (volume 625, 2024; DOI 10.1038/s41586-023-06924-6). FunSearch produced genuinely new, record-beating solutions on a long unsolved problem; the problem itself remains open.
Going deeper
Section titled “Going deeper”- Official HKS course listing for DPI-681M, Harvard Kennedy School. The in-person course behind the open online materials this track adapts.
On this site
Section titled “On this site”- What generative AI actually is, lesson 1 of this track. The prediction-machine insight that this lesson’s context and variability sections build on.
- How AI reads your words (Transformers and LLMs track, shown as AI Foundations in the sidebar). The machinery underneath the ranked menu of next words, for readers who want to go a level down.
Adjacent topics
Section titled “Adjacent topics”- The next lesson in this track steps beyond the chat window: why some assistants arrive already knowing their job before you type a word, how an AI can work from your own documents, and what changes when these systems can see images and speak out loud.
- The lesson’s caution that a confident persona can still be wrong on facts returns in the track’s second arc, where the later lessons take up risks and what to do about them.