Example

Podcast 2 — In conversation

Speaker A: Could you give me an example of how to use this word in a sentence? That's literally the most common way we use example — asking for one.

Speaker B: True. And that request itself is an example of example in action. The word is almost self-referencing — you need examples to explain what an example is.

Speaker A: So the primary meaning is a specific instance that illustrates a general point?

Speaker B: Exactly. If I say fruits are healthy and then mention apples and oranges, those are examples of fruits. They represent the larger category.

Speaker A: But there's also the moral meaning, isn't there? Like she leads by example.

Speaker B: Right. That's the second major sense — a person or behaviour worth imitating. To set an example means to behave in a way others should follow.

Speaker A: And then the negative flip — to make an example of someone means to punish them publicly as a warning.

Speaker B: Yes, quite a dark usage. The judge made an example of the offender means the punishment was deliberately harsh to deter others.

Speaker A: What synonyms come close to example? I'm thinking of instance, illustration, specimen.

Speaker B: Good list. Instance is the closest — it's almost interchangeable in many contexts. For instance and for example mean the same thing. But illustration implies visual or explanatory clarity, and specimen has a scientific, physical connotation.

Speaker A: So instance is neutral and general, illustration emphasises explanation, and specimen implies something you can examine physically?

Speaker B: Precisely. Example covers all those nuances and more — it's the broadest and most versatile of the group.

Speaker A: A common mistake — people sometimes write for example with a comma but forget the comma after the examples themselves.

Speaker B: That's a punctuation issue more than a vocabulary one, but it's worth noting. For example, apples and oranges — always a comma after example in that construction. And the abbreviation is e.g., from the Latin exempli gratia, meaning for the sake of example.

Podcast 3 — Prompt engineering

Speaker B: Today we're looking at example in development prompts. This word is arguably the most powerful prompt engineering tool in existence — because AI models learn best from examples.

Speaker A: That makes sense. Few-shot prompting is literally providing examples for the model to follow.

Speaker B: Exactly. Let's start with a UI component prompt:

Speaker B: PROMPT: Create a reusable form validation library and include a working example for each validator type. Provide an example login form with email validation, an example registration form with password strength checking, and an example contact form with phone number formatting — each example fully functional and copy-pasteable.

Speaker B: Including example for each tells the AI not just to build the library, but to demonstrate it. The AI will generate documentation with runnable code because you asked for examples explicitly.

Speaker A: That's much better than just saying build a validation library. What about database design?

Speaker B: Here's one:

Speaker B: PROMPT: Design a multi-tenant database schema for a SaaS invoicing platform. Include example seed data with three example tenants, example invoices in various states like draft, sent, paid, and overdue, and example queries showing how to retrieve all overdue invoices for a specific tenant with pagination.

Speaker B: Example seed data and example queries tell the AI to go beyond schema design and provide realistic test data plus working SQL. The AI generates a complete runnable database, not just table definitions.

Speaker A: That transforms a design task into a usable deliverable. What about API development?

Speaker B: Here's a REST API prompt:

Speaker B: PROMPT: Build a RESTful API for a task management system with full CRUD operations. For every endpoint, include an example request with headers and JSON body, an example success response, and an example error response showing proper error codes and messages. Format all examples as curl commands.

Speaker A: That's brilliant — the AI will generate complete API documentation with copy-pasteable curl commands. What about application architecture?

Speaker B: Good question:

Speaker B: PROMPT: Create a complete role-based access control system for a healthcare application. Include an example role hierarchy with administrator, doctor, nurse, and receptionist roles, example permission matrices showing which roles can access which patient data, and example middleware code demonstrating how to enforce permissions at the route level.

Speaker B: Example role hierarchy and example permission matrices give the AI concrete data to model. Without those words, you'd get abstract descriptions instead of working code.

Speaker A: The difference between theory and practice — examples force the AI to be practical. What about testing?

Speaker B: Here's one for test automation:

Speaker B: PROMPT: Design a test automation framework for an e-commerce checkout flow. Include example test cases for successful purchase, failed payment, coupon application, and cart abandonment. Provide example page objects for the cart page and checkout page, and example test data fixtures with realistic products, prices, and user profiles.

Speaker A: That gives the AI a full testing blueprint — page objects, fixtures, and test cases all anchored by the word example.

Speaker B: And one final comprehensive prompt:

Speaker B: PROMPT: Build a complete design system component library with a living style guide. For every component — buttons, modals, tables, forms, navigation — provide an example in each variant such as primary, secondary, danger, and disabled states. Include an example page layout combining multiple components, example responsive breakpoints, and example dark mode and light mode theme tokens with before-and-after example screenshots described in alt text.

Speaker A: That's a full design system spec where example appears at every level — component, variant, layout, theme. The AI builds a complete reference guide because every section demands concrete examples.

Speaker B: Exactly. In prompts, example is the most reliable word for transforming abstract instructions into concrete, usable output. It forces the AI out of theory and into practice. Every time you add include an example, you're telling the AI — show me, don't tell me. That single word consistently produces higher-quality, more actionable code.