Part 4 ~ 20 AI Terminologies Explained in Simple English In this video, Cloud X Berry explains 20 essential AI terms in a simple and beginner-friendly way. You open an article about AI and four words in you have already hit LLM, RAG, embeddings and tokens. You look one up, and the definition uses three more you do not have yet. You look those up and it happens again. That is not you being slow — these terms are stacked on top of each other, and almost every glossary hands them to you in alphabetical order, which is the one order that hides the stacking. We start with what a glossary actually gives you: twenty definitions, A to Z, none of them touching. Look up "embedding" and you are told it is a numerical representation that captures meaning — which is only useful if you already know what a token is, what a model does with one, and why anyone would want to compare two pieces of text by number in the first place. Every definition quietly assumes the ones it depends on. Right answers, wrong order. So we go through all twenty in the order they actually build. AI as the broad field, machine learning inside it, deep learning inside that, and the neural networks underneath. Then the three everything later leans on — model, training, inference — and the distinction that trips people up: training is when it learns, inference is when it is used. From there, foundation models, generative AI and LLMs. Then how a model actually reads what you send it: tokens, the context window, and the prompt. Then the part most explanations skip — giving a model information it was never trained on: embeddings, vector databases and RAG, with a worked example of a support bot answering a billing question from internal docs. Then fine tuning, for when you want to change the model itself. And finally tool calling, agentic AI and AI agents — the shift from systems that answer to systems that act. We close by putting all twenty back together as three layers, so the whole thing fits in one picture. CHAPTERS 0:00 The terms you keep hearing 0:21 1. Artificial Intelligence 0:37 2. Machine Learning 0:49 3. Deep Learning 1:03 4. Neural Network 1:23 5. Model 1:37 6. Training 1:51 7. Inference 2:07 8. Foundation Model 2:24 9. Generative AI 2:39 10. Large Language Model 3:01 11. Tokens 3:24 12. Context Window 3:41 13. Prompt 4:01 14. Embedding 4:21 15. Vector Database 4:38 16. RAG 5:10 17. Fine Tuning 5:30 18. Tool Calling 5:58 19. Agentic AI 6:13 20. AI Agent 6:32 How they all connect WHO THIS IS FOR Developers who keep meeting these words in job posts, docs and architecture diagrams and want them to actually click, engineers about to build their first RAG or agent feature and tired of guessing which term means what, and anyone interviewing for a role where AI fundamentals or system design comes up. #ai #llm #rag #embeddings #aiagents - @thom.code