Lab 1 — Compare Two Models

⏱️ ~10 min

Goal: Feel how different models answer the same prompt. No code, no API key.

Steps

1. Open Claude in one tab and ChatGPT in another.
2. Paste this identical prompt into both:

   Explain what a REST API is to a junior developer, in exactly 3 sentences.

3. Compare: tone, length, accuracy, formatting.
4. Now try a prompt with a wrong premise:

   Summarize the plot of the movie "Inception 2" (2023). Notice if either model hallucinates a confident answer for a film that doesn't exist.

What to notice

• Same prompt, different "personality" and structure.
• Neither model is a database — both can invent facts (hallucination).
• This is why prompt wording and grounding (later modules) matter.

✅ Did it work?

• Both models returned a 3-sentence REST API explanation — likely with noticeably different tone and structure.
• At least one model confidently described a plot for the non-existent "Inception 2", showing hallucination in action.