Turn scattered AI knowledgeinto workflows you can run.
8 structured courses laid out as one path, foundational to advanced: from working out what a token actually is, to handing a whole task to an agent.
8 coursesDeeper color, further along
01Starting from zero16 lessons
AI Foundations
Beginners who want a clear understanding of core AI concepts, capability boundaries, and basic safety before using AI at work.
Understand LLMs, tokens, context windows, hallucination, privacy boundaries, embeddings, multimodality, reasoning models, agents, model size, and open vs. closed models — and use them to make real tool decisions.
01What an LLM Actually Is: A Very Good Guessing Machine
02How It Was Trained: Pre-training, Fine-tuning, and RLHF
03Tokens: What AI Measures In, and What You Pay In
People already using ChatGPT, Claude, or Gemini who get inconsistent answers and want a reliable way to state what they actually need.
Master CO-STAR, chain of thought, few-shot examples, and constraint lists; control output formats; and build your own AI assistant with system prompts.
Professionals ready to put AI into daily work but not yet settled on a product — and anyone already using one who wants the underlying method.
Judge which category of tool a task calls for, write and reuse prompts and system prompts that hold up, apply working methods that survive a change of product, and cover images, voice, video, AI search, knowledge bases, and meeting assistants.
Professionals making Gemini their primary tool — especially anyone already living in Gmail, Docs, Sheets, and Drive.
Complete sign-up and plan selection, master the machinery that is Gemini's alone (Gems, NotebookLM, the Chrome sidebar), and handle email, data, presentations, writing, and long documents across Gmail, Sheets, Slides, Docs, and Drive.
01Gemini 实用指南：如何用好 Google 生态的 AI 神器
02Free / AI Plus / AI Pro / AI Ultra 与 Google Workspace 的关系
People who are already fluent with chatbots and want to take their first step into agents, then learn to use and create Skills.
Understand how agents and Skills relate, install and judge existing Skills, and have an agent help you write your first Skill that triggers reliably and passes its tests.
Advanced users who are already fluent in conversation with Claude but have never had it deliver a finished piece of work — people who want to hand off the filing, research, weekly reports and follow-ups that eat their day.
Draw the line between Cowork and Chat, run the full authorize–plan–approve–accept loop, connect folders, connectors and the browser, and use five worked cases to turn everyday work into tasks that can be delivered again and again.
Every course here is written in-house as a structured curriculum, not a link roundup. Lesson counts are the real chapters published on this site for this language, appendices included.
Courses open in batches — do not miss the launch note
Subscribe and new courses and updates reach your inbox along with the Brief.