📖 Plain-English AI glossary
AI terms, explained for total beginners
New to AI and drowning in jargon? Here are the words you keep seeing — LLM, prompt, hallucination, token, RAG, agent — each in one or two clear sentences, no math and no hype.
Every term links to the free bloom0 lesson where you actually use it. Free forever · no sign-up to start · also in Español.
The words, from first idea to real building
- Artificial Intelligence (AI)
- Software that does things we used to think needed a human mind: writing, answering questions, making images, or spotting patterns. Today's popular AI tools are mostly "generative" — they create text, code, or images on demand. Start in lesson 1 →
- Large Language Model (LLM)
- The engine behind chatbots like ChatGPT, Claude, and Gemini. It's trained on huge amounts of text and works by predicting the most likely next word, one piece at a time — which, at massive scale, produces answers that feel intelligent. Lesson 1 →
- Prompt
- The instruction or question you type to an AI. A clear prompt with context, a specific task, and the format you want gets far better results than a vague one. Lesson 3 →
- Prompt engineering
- The skill of writing prompts that reliably get good results — using patterns like giving the AI a role, showing examples, or asking it to critique its own answer before finalizing. Prompting cheat sheet →
- Hallucination
- When an AI gives an answer that sounds confident but is actually wrong or made up. It happens because the model predicts plausible text, not verified facts — so always check important claims. Lesson 1 →
- Token
- The small chunk of text — roughly three-quarters of a word — that an LLM reads and writes in. AI tools measure length and cost in tokens, so "how many tokens" really means "how much text." Apps track →
- Context window
- How much text an AI can hold in mind at once — your prompt plus its reply, measured in tokens. Go past it and the model starts forgetting the earliest parts of the conversation. Advanced AI track →
- System prompt
- Hidden instructions that set an AI's role, rules, and tone before your conversation starts, for example "You are a friendly cooking assistant." It quietly shapes every reply. Lesson 20 →
- Chatbot (AI assistant)
- A conversational AI you talk with in plain language, like ChatGPT, Claude, or Gemini. Treat it as a partner you can ask follow-ups and push back on, not a search box. Lesson 2 →
- Model
- One specific trained AI you can use, with a name and version like GPT-4o or Claude Sonnet. Bigger or newer models are usually more capable but can cost more or run slower. AI tools cheat sheet →
- Training data
- The huge collection of text and images an AI learned its patterns from. It's why a model can write in many styles — and why it can be out of date or biased, reflecting whatever it was trained on. Lesson 1 →
- Fine-tuning
- Further training a base model on your own examples so it specializes in a task or voice. Most beginners never need it — good prompting and RAG solve most problems first. Advanced AI track →
- API
- A doorway that lets your own app send text to an AI and get a response back in code, instead of typing in a chat window. It's how you build AI into a product. Lesson 18 →
- RAG (Retrieval-Augmented Generation)
- Giving an AI your own documents to answer from, so it responds using your knowledge instead of guessing. The app finds the relevant passages and hands them to the model along with your question. Lesson 25 →
- Embedding
- A way of turning text into a list of numbers that captures its meaning, so a computer can measure how similar two pieces of text are. It's the trick that powers semantic search and RAG. Lesson 26 →
- Vector database
- A store built to hold embeddings and instantly find the most similar ones. It's the searchable "memory" a RAG app pulls the right context from. Lesson 27 →
- AI agent
- An AI that doesn't just answer — it takes actions in a loop: it can use tools, run steps, and work toward a goal with less hand-holding from you. AI agents cheat sheet →
- Multimodal
- AI that handles more than text — images, audio, or voice — in the same model, so you can show it a photo or talk to it instead of only typing. Lesson 29 →
- Vibe coding
- Building software by describing what you want in plain language and letting an AI write and fix the code with you, instead of memorizing syntax. Lesson 16 →
- Open weights (open-source AI)
- Models whose files are published so anyone can download and run them, like Llama or Mistral — versus closed models you can only reach through a company's paid service. Advanced AI track →
Now go build something
You don't learn AI by reading definitions — you learn it by making things. Every term above shows up in a hands-on bloom0 lesson where you build the real thing.
Explore the tracks
Free printable cheat sheets