Post-Lab 2 Activity

You will answer these questions and submit by Friday evening. Answer Q1, Q2, and Q3, then additionally pick at least two of Q4-6 to complete. For Questions 3-6, you may share the same AI use attempt with your group members. However, all group members should directly observe inputs and outputs, and each member should write their own answers. If you work in a group, you are welcome to collaborate directly or work separately on items and compare observations with each other. Answers should be written by human students! Again, “raw” observations in incomplete sentences or with nonstandard grammar are welcome.

When you are finished, write up your collaboration statement (Question 7).

1. (You may submit text identical to your group’s for this question) Please describe what you were trying to accomplish in Lab 2 Part II. Additionally, copy and paste your list of blockers from Lab 2 Part II
2. (Please come up with individual answers) Before you try anything: what do you expect to see when you try to use AI for your problem? What things do you expect that AI tools will be helpful for? What things do you expect your group’s human judgment to be useful for? What would you want to do or check before accepting an AI output?
3. Pick a GenAI tool to use and name it. Try to prompt it to complete your task end-to-end, all in one go, modulo follow-up questions it asks you. Paste your input in your response here! What happened? What worked well, and what didn’t? Did its attempt surprise you in any way (e.g. its approach, solution, how long it took)?
4. Use the same tool as in Q3. Use a similar prompt, but this time, also describe your roadblocks and provide them as context. Again, paste your input into your response for this question, and describe your observations (What happened? What worked well, and what didn’t? Did its attempt surprise you in any way?)
5. Use the same tool as in Q3. Now, instead of prompting the tool to complete your Lab 2 Part II problem end-to-end, identify a specific component of it (e.g. “just” output validation, a loop to collect comments in a thread, or even a regex that you can describe precisely but would rather not write by hand). Describe the component you choose, and again provide your input and describe your observations (What happened? What worked well, and what didn’t? Did its attempt surprise you in any way?)
6. Pick either the original end-to-end scope or your refined scope from Q5 (if you completed Q5). Choose a different tool, try using it, and note your observations again. What tool did you use? What happened? What worked well, and what didn’t? Did its attempt surprise you in any way?
7. (You may submit text identical to your group’s for this question) Did you work with anyone? If so, state their name(s) and Andrew ID, and briefly describe how you worked together– whether you pair programmed, each tried different approaches, at what point(s) you came together to discuss your observations, etc.

Note: We expect that some students will be constrained by their own tool access! For example, one group member might subscribe to Claude Code, while another does not– in this case, you may want to either compare approaches or share one person’s setup for this activity. Moving forward, we will be providing some tools!