Part 3/9:
Constructing Effective Prompts: A Step-by-Step Approach
Shapiro illustrates prompt engineering through real-world examples, starting with a simple request: generate vivid, romantic descriptions of wines for an online store. Initially, he primes ChatGPT with a context setting—clarifying that it should understand the concept of prompt engineering—and then constructs a detailed, example-driven prompt for wine descriptions.
He emphasizes that less is often more: a concise, well-structured prompt tends to produce better results. For instance, instead of a lengthy instruction, a streamlined prompt like "Create a vivid wine description from the perspective of a French sommelier, including flavor profile and bouquet" can yield impressive, engaging descriptions.