Prompt Engineering

The ability to get useful outputs from AI — called prompt engineering — is increasingly a practical skill. More importantly, it reveals a lot about how AI thinks.

Why prompting matters

Two people asking the same AI the same question can get dramatically different results depending on how they phrase it. This is because AI responses are extremely sensitive to the phrasing, framing, and context of the prompt. Learning to prompt well is learning to communicate clearly with a very particular kind of system.

Key prompting techniques

Role assignment. Telling the AI to act as a particular expert changes its register and depth: "You are a marine biologist. Explain why coral bleaching matters."

Chain of thought. Asking AI to "think step by step" or "show your reasoning" before giving an answer improves accuracy on complex reasoning tasks — because it forces the model to generate intermediate reasoning tokens rather than jumping to a conclusion.

Constraints and format. Specifying exactly what you want: "In three paragraphs, no more than 200 words total, using language appropriate for a 14-year-old."

Examples. Providing one or two examples of the format or style you want dramatically improves outputs for structured tasks.

What prompting reveals

Experimenting with prompts reveals interesting things about how LLMs work: they are sensitive to framing in ways humans often are not, they can be led to contradict themselves, and their "reasoning" is not as robust as it looks. This knowledge is useful both for using AI effectively and for understanding it critically.

Good prompting is not about gaming the system — it is about communicating clearly with a tool that requires specificity to perform well.