Zero-shot means asking a model to do a task with no examples — just the instruction. Few-shot means providing a handful of input-output examples in the prompt before the actual request. "Here are 3 examples of how to format this data... now do this one." The model learns the pattern from context alone, no training required.
Why it matters
Few-shot prompting is the fastest way to teach a model a new format or behavior. Need consistent JSON output? Show it three examples. Need a specific writing style? Give it samples. It's free, instant, and surprisingly powerful.