AI Jargon Decoder
LLM, MCP, RAG, hallucination, fine-tuning, tokens — what AI terms actually mean in plain English for busy professionals.
AI Jargon Decoder
AI conversations are full of acronyms and technical terms that sound intimidating but aren't that complicated. Here's what they actually mean.
LLM (Large Language Model)
The technology behind ChatGPT, Claude, Gemini, etc. Think of it as a very sophisticated autocomplete that's read most of the internet. It predicts the next word in a sequence, which turns out to be surprisingly useful.
MCP (Model Context Protocol)
A standard way for AI tools to connect to your existing software. Think of it as USB-C for AI. Before MCP, every AI tool needed its own custom integration. Now there's one standard plug.
RAG (Retrieval-Augmented Generation)
When an AI looks up specific documents before answering your question, instead of relying only on its training. It's like the difference between answering from memory vs. checking your notes first.
Hallucination
When AI confidently states something that's completely false. Not a bug that will be fixed. It's a fundamental feature of how these systems work. Always verify important claims.
Fine-tuning
Training an AI model on your specific data so it performs better for your use case. Like teaching a general contractor the specific building codes for your city.
Token
How AI measures text. Roughly 1 token equals 0.75 words. When people talk about "context windows" or "token limits," they mean how much text the AI can process at once. Bigger is better.
Context Window
The AI's working memory. How much text it can "see" at once during a conversation. Think of it as the size of the desk. A bigger desk means more documents open at the same time.
Prompt Engineering
The art of writing instructions that get AI to do what you want. Sounds fancy, but it's mostly just being clear and specific. See our prompting guide for the practical version.
Vibe Coding
Using AI to write software by describing what you want in plain English instead of writing code. Real, but early. Non-developers can now build simple prototypes. Complex software still needs real engineers.
Agentic AI
AI that can take multi-step actions autonomously instead of just answering one question at a time. The buzzword of 2026. Real technology, overhyped timelines. See our AI agents guide for the full breakdown.
Rule of thumb: If someone uses a term you don't know, ask "What does that mean for my work this week?" If they can't answer in one sentence, it probably doesn't matter yet.
Get new guides delivered every Tuesday.
AI news, prompts, and workflows you can use between meetings. Under 60 seconds.