Learn AI
like a
Doctor

A model is a large neural network trained on billions of words from the internet, books, and scientific papers. It learns patterns in language so well that it can answer questions, write documents, summarize data, and reason through complex problems. Think of it like a very well-read consultant available 24/7.

Tokens are the pieces AI splits text into before reading it. A token is roughly 4 characters or 3/4 of a word. The sentence 'Hello, how are you?' is about 5 tokens. AI models have limits on how many tokens they can process at once — this is the context window.

A prompt is any text you send to an AI model to get a response. Great prompts have four parts: Role (who the AI should be), Task (what to do), Context (background information), and Format (how to structure the output). Better prompts produce dramatically better results.

The context window is the maximum amount of text (measured in tokens) an AI can process in a single interaction. It includes your prompt, the conversation history, and the AI's response. When the context fills up, older messages are forgotten. Modern models have windows from 8K to 200K tokens.

Embeddings convert text into lists of numbers (vectors) that represent meaning. Words or sentences with similar meanings end up close together in this mathematical space. This allows AI to find related content even when the exact words don't match — powering semantic search and retrieval systems.

By default, AI has no memory between conversations. But memory can be added in four ways: no memory (every session is fresh), session memory (remembers within one conversation), long-term memory (stores and retrieves past information), and tool-backed memory (reads from databases, EHRs, or files).

LLM stands for Large Language Model. These are AI systems trained on massive amounts of text using transformer architecture. They predict the most likely next token given previous tokens. Despite this simple mechanism, LLMs demonstrate remarkable reasoning, coding, writing, and analysis abilities.

A chatbot takes a question and returns an answer — one step, then done. An agent can plan multiple steps, use tools (search, database, calculator), loop back if something goes wrong, and complete a complex task autonomously. Agents are chatbots with superpowers: memory, tools, and the ability to take actions in the world.

An AI agent is a system that perceives its environment, reasons about goals, plans a sequence of actions, executes tools, observes results, and adapts until the task is complete. Unlike chatbots, agents can take actions: search the web, query a database, send emails, fill forms, or call APIs.

Tools are functions that an AI agent can invoke to take actions beyond just generating text. Common tools include: web search, database queries, calculator, file reader, calendar API, email sender, and form filler. Tools are what transform a chatbot into an agent capable of completing real tasks.

Skills are pre-built, tested prompt templates that an agent can deploy for specific tasks. A skill bundles together the role, instructions, format, and examples needed for one job. Think of skills as the AI's job description cards — each one tells it exactly how to handle a particular situation.

A workflow is a defined sequence of steps: what data to gather, which tools to run, which agent handles each stage, what to do if something fails, and what the final output looks like. Workflows are written in plain language or code and are the SOPs of your AI system — they make AI behavior predictable and auditable.