智能体重要知识点速览

Agent Essentials: A Quick Tour

Ten short chapters to grok every word of the agent era: LLMs, foundation models, multimodality, tokens, hallucination, agents, MCP, skills

8 labs10 AI-mentored sessions~2 hoursBilingual · EN / 中
Start learning →

About this course

You hear these words every day: large models, foundation models, multimodality, context
windows, hallucination, agents, MCP, skills, RAG. What are they, really? This course writes
no code. It explains each word through small experiments you can see and drag — you'll click
out an LLM's "next word", drag a context window to watch it forget, deliberately make it
invent a falsehood, plug tools into an agent, and install a skill. Eight to fifteen minutes
per chapter, with an AI mentor you can interrogate at any point. You won't become an engineer;
you will genuinely understand this era — and know where to trust it and where to doubt it.

What you'll learn

  • Explain, in plain language, what LLMs, foundation models and multimodal models are, and how they relate
  • Explain how tokens and the context window decide what an AI can read, how it's billed, and what it forgets
  • Explain why AI confidently makes things up, and when to doubt it
  • Tell apart when to use prompting versus fine-tuning
  • Explain how an agent gets work done via think → act → observe, and how RAG lets it answer open-book
  • Tell apart MCP (plugging in tools) from Skills (teaching it procedures), and know how to extend an agent

Syllabus

1Foundation Models & LLMs: What's the Brain Underneath an Agent6 sessions
  • 1What Is a Large Language Model: The Internet's Super-Autocomplete12 minStart →
  • 2What Is a Foundation Model: One Base, Countless Applications12 minStart →
  • 3Vision Models & Multimodality: Giving AI Eyes and Ears12 minStart →
  • 4Tokens & the Context Window: What You're Billed For, and What It Forgets12 minStart →
  • 5Probability & Hallucination: Why It Lies with a Straight Face12 minStart →
  • 6Prompting vs Fine-Tuning: Brief Them, or Send Them to School12 minStart →
2Agents, Tools & Skills: How an Agent Actually Gets Work Done4 sessions
  • 1What Is RAG: Let AI Take the Exam Open-Book12 minStart →
  • 2What Is an Agent: Giving the Model Hands and Feet13 minStart →
  • 3What Is MCP: The Universal Socket for Plugging Tools into AI12 minStart →
  • 4What Is a Skill: Teaching an Agent How You Do Things13 minStart →
3Prompting & Context: Saying What You Want0 sessions

Sessions are on the way.

4Retrieval (RAG): Feeding AI Your Own Docs0 sessions

Sessions are on the way.

5Memory & State: How AI Remembers0 sessions

Sessions are on the way.

6Multi-Agent: Teams of AIs0 sessions

Sessions are on the way.

7Evaluation & Trust: Is the AI Any Good?0 sessions

Sessions are on the way.

8Safety & Alignment: Keeping AI in Check0 sessions

Sessions are on the way.

Same series · 面向非科班的快速能量补充

Fast Track for Non-CS Backgrounds

Not from a CS background? Get up to speed fast — the essentials of agents and computer science.