Why You Need to Know These Words (Or: How to Sound Smart at Parties)
You don't need to be a programmer to use AI. But knowing 10 key terms turns you from someone who "tries AI" into someone who commands it.
Think of this like learning to drive. You don't need to know how an engine works. But you should know what the pedals do. Also, knowing the words makes you sound terrifyingly intelligent in meetings. Use "context window" three times in one sentence and watch people nod respectfully.
---
The 10 Words You Actually Need
| Term | Plain English | Why It Matters |
|------|--------------|----------------|
| LLM | A giant brain trained on most of the internet. ChatGPT, Claude, and Gemini are all LLMs. | Every AI tool you'll use runs on an LLM. They're the engine. |
| Prompt | The message you type into the AI. That's it. A question, an instruction, a request. | Better prompts = better results. The whole point of this course. |
| Token | A chunk of text — roughly ¾ of a word. "Hello world" = 2 tokens. | Determines how much you can type and how long the AI's response can be. |
| Context Window | How much the AI can "remember" in one conversation. Like its short-term memory. | Bigger context = the AI can handle entire books, not just paragraphs. |
| Hallucination | When the AI confidently says something completely wrong. It sounds real but isn't. Like that coworker who nods along in meetings and then says something wildly incorrect with total conviction. | Always fact-check AI outputs. It's not lying — it's guessing confidently. |
| Training | The process of teaching an AI by showing it millions of examples. Like showing a child 10,000 pictures of dogs until they finally stop calling everything with four legs a "dog." | Happens once, before you ever use it. You don't need to do this. |
| Inference | When the AI uses its training to answer your question. | Every time you type a prompt, you're running inference. That's the magic. |
| Fine-tuning | Taking a pre-trained AI and teaching it extra stuff (like your company's writing style). | Advanced. Most people never need this. Useful for specialised tasks. |
| Temperature | How "creative" the AI is allowed to be. Low = predictable. High = surprising. | Low (0.1) for facts. High (0.8) for creative writing. Default is usually fine. |
| API | A way for programs to talk to each other. Lets you plug AI into your own apps. | Advanced. But this is how businesses automate with AI. |
---
🧠 A Quick Way to Sound Like You Know What You're Talking About
| Don't Say | Say Instead |
|-----------|-------------|
| "The AI chat thing" | "I'm using an LLM" |
| "It made stuff up" | "It hallucinated" |
| "I ran out of room" | "I hit the context limit" |
| "How creative is it?" | "What temperature setting?" |
| "Tell me about..." | "Here's my prompt..." |
---
🛠️ 30-Second Practice
Open ChatGPT or Claude right now. Type this:
"Explain the term 'context window' using a sandwich analogy. Keep it under 3 sentences."
See? Now you already know enough to test whether the AI understands the concept correctly. That's the power of knowing the vocabulary.
---
What We've Covered
✅ LLM — the engine behind every AI tool
✅ Prompt — the message you send
✅ Token — how AI measures text
✅ Context window — AI's short-term memory
✅ Hallucination — when AI is wrong but sounds right
✅ Temperature — how creative the AI gets