Compound

Podcast 2 — In conversation

Speaker A: I've been looking into compound interest for my savings. It's incredible how fast your money grows over time.

Speaker B: Tell me about it. I started investing in an ISA and the compound growth is really noticeable after a few years.

Speaker A: Right, that's the magic — you earn interest on your interest. The returns compound over time.

Speaker B: Exactly. And it's not just money. Knowledge compounds too. The more you learn, the faster you can learn new things.

Speaker A: That's a good point. Like when you're learning a language — once you know the basics, new vocabulary just clicks faster.

Speaker B: Oh, completely. Compound learning is a real thing. Each concept you master makes the next one easier to absorb.

Speaker A: What about compound words? My daughter is learning about them in school.

Speaker B: Ah yes — words like 'butterfly' or 'raincoat'. Two separate words joined together to make something new.

Speaker A: And the meaning isn't always obvious from the parts. 'Hotdog' has nothing to do with dogs!

Speaker B: True! That's what makes compound words interesting. The whole becomes more than just the sum of its parts.

Podcast 3 — Prompt engineering

Speaker B: Welcome back. Today we're exploring how 'compound' helps us craft better prompts for AI, especially when dealing with complex data structures and financial calculations.

Speaker A: I use 'combine' a lot in prompts. What's the difference with 'compound'?

Speaker B: Great question. 'Compound' suggests accumulation and growth over time, while 'combine' is more static. Here's our first example.

Prompt: Build a compound interest calculator that compounds monthly contributions with yearly raises factored in. Show growth projections over 5, 10, and 20 year periods.

Speaker B: Notice how 'compound' here signals the recursive nature of interest calculations — each period building on the previous one.

Speaker A: So it's about iterative growth. What about data structures?

Speaker B: Here's a data structure example. Prompt two.

Prompt: Design a compound data structure for a portfolio tracker. Combine stock holdings, bond positions, and cash reserves into a unified risk assessment model.

Speaker B: 'Compound data structure' tells the AI we need interconnected, layered data with multiple components working together.

Speaker A: I see. What about CSS styling?

Speaker B: Here's a styling example. Prompt three.

Prompt: Create compound CSS selectors combining pseudo-classes and attribute selectors. Build a theme system with primary, secondary, and compound color variants.

Speaker B: This guides the AI toward creating layered, composable styles that build on each other.

Speaker A: What about building search functionality?

Speaker B: Here's a search example. Prompt four.

Prompt: Implement a compound search engine that combines full-text search, filters, and relevance scoring. Support AND, OR, and NOT boolean operators.

Speaker B: 'Compound search' signals we want multiple search mechanisms working together, not just keyword matching.

Speaker A: And for authentication?

Speaker B: Here's a security example. Prompt five.

Prompt: Design a compound authentication system layering password, TOTP, and biometric verification. Include fallback recovery options for each layer.

Speaker B: This shows how multiple security mechanisms compound to create a robust authentication flow.

Speaker A: 'Compound' really does suggest the right architectural thinking.

Speaker B: It does. 'Compound' signals accumulation, layering, and recursive growth. Use it when building systems where each component builds on the previous one.