From entry to workflow: Coach Jiang’s AI course map | Coach Jiang

What is this article talking about?

The most common way to get lost when learning AI is to treat each class as a single tool presentation: learn the presentation today, learn the verbatim draft tomorrow, and still don’t know the next step after learning. The function of the course map is to answer "what to learn next": connect each class to a competency node and connect the things learned to each other. This article lays out the map of my entire course. Who is

suitable for?

· People who have taken a few AI classes and want to know where they are now and where to go next.
· People who want to systematically move from getting started with tools to personal work systems
· Teachers and knowledge workers who are designing their own curriculum systems can refer to this map logic

What can you take with you?

· The structure of the entire course map: foundation layer, apply upwards, and dig deeper
· The content of two layers of each of the three deep excavation routes, and their final destination
· The connection logic between courses: at the end of each step, questions for the next step naturally arise

Underlying cognitionThe content taught in this map will not expire as long as AI remains a Large Language Model (LLM) architecture. Tools will change layouts, names, and concepts such as word relevance, context, and tacit knowledge. The more advanced AI becomes, the more useful it becomes.

Shape of the map: one base, two directions

The entire map is a vertical structure. There is a foundation layer in the middle, which is the starting point for everyone: the basic course of Agent introduction, which teaches the most basic content of setting up the AI Agent working environment, software installation, environment construction, and core concepts. After taking this class, you will get a basic Agent environment that can actually operate. I call it "AI Manager Youth Edition". This is the prerequisite for all subsequent routes.

Start from the foundation layer and expand in one direction upwards and downwards:

Up: Apply other people’s tools

Use the prepared skill packages and workflows directly to solve immediate problems quickly without the need for in-depth study. For example, meeting minutes workflow and sustainability report workflow can be run as soon as you take them. Most of them areSkill Pack Download Page You can get it with.

Down: Dig deeper into your own system

Through long-term accumulation of courses and coaching services, study the parts you are good at and build a good system. Slow, but it will compound.

The two roads do not conflict. If you are good at knowledge management but not good at sustainability reporting, dig deeper to refine the tacit knowledge, and at the same time directly apply the sustainability reporting workflow upwards. If you are good at it, dig deep on your own, and if you are not good at it, use other people's methods, and go at the same time.

Dig deeper: three parallel routes, each digging two levels

There are three parallel routes in the direction of deep excavation, extending directly from the foundation layer to the next layer below:

Left line・First floorData organization technology in the AI era

Use AI to organize data so that the knowledge base can be understood and used by AI. In addition to learning tools, the data collection class also teaches how to use data by AI.

Middle line・First floorTacit knowledge extraction basics

To turn one-time experiences into reusable knowledge assets, first dig out the things that "can be done but cannot be explained".

Right line・First floorBasics of Semantic Engineering

Understand the semantic space and word correlation, and learn to navigate in the probabilistic world of AI. In addition to learning sentence patterns, the prompt word class also teaches how to explain tasks clearly.

Left line・Second floorAgent General Manager Personal Management System

Upgrade the AI manager from the youth version to the full version and establish a personal management system.

Middle line・Second floorTacit knowledge refining advanced

Dig deeper judgment standards from the emotional entrance and establish your own thinking framework.

Right line・Second floorFour-dimensional label management

Create a controlled vocabulary using the four dimensions of object, content, context, and nature, allowing AI to accurately understand your classification logic.

The three lines finally converge to the same end point: Agent, a one-person company operations team. Use an AI Agent team to support the operation of a one-person company. The data collection, tacit knowledge, semantic engineering, and management system learned earlier are all connected here.

How to connect the courses: Let the next questions grow on their own

The courses on the map have a common design principle: at the end of each step, students will naturally generate questions for the next step, and there is no need to push them.

Free general education lecture on "Word Relevance and Prompt Word Design". The core is that communication is two-way. How you speak, how AI listens. People who have finished learning usually ask: "Now that the prompt words have been designed, how can they be used other than posting dialog boxes?" This leads to the next group.

"Attention Mechanism and Contextual Engineering" answers this question and talks about how AI remembers and how you manage it: AI has no memory, only windows, dialog boxes, project modes, and skill packs. The differences and timing of use of these containers. People who have finished the course will ask: "Containers are used, but what is the most important thing to put in them?"

The answer is tacit knowledge refining: the raw material for AI operation is your knowledge, and the most valuable ones are usually those that you can do but can’t explain clearly. The workshop will take you to actually dig one or two layers, and then divide into two directions: dig deeper into the underlying logic and judgment criteria, or expand horizontally to rule base design. The knowledge base allows AI to empower the brain (understand your expertise), and the rule base allows AI to empower the hands and feet (execute for you). When the two are strung together, it goes from "AI understands you" to "AI does things for you."

The sequence has a prerequisiteDigging deep is the prerequisite for well-written rules. The rules must be written clearly, but where do they come from? Refined from a deeper level. The deeper it is refined, the more stable the underlying logic is, and the more the written rules can cope with new situations. There is no need to make up a rule with each case.

How you can use this map: Find your location first

The first step when looking at a map is positioning. First determine which stage you are at, and then the next class will become part of the route.

If there is no Agent environment yet: Before you build the foundation layer, set up the environment first.
If you know how to use tools but the data is messy: take the left route and organize the data first so that the knowledge base can be used by AI.
If you have the information but can’t explain your major clearly: take the middle line and refine tacit knowledge.
Commands often misunderstood by AI: take the right route, semantic engineering.
There is an urgent matter to be solved: go up and directly implement the workflow, and follow the other path.

After knowing the position, each lesson will be connected to your long-term ability, and learning will change from a collection tool to a accumulation system. I want to start from my own end first.Use AI to build a traceable learning map That articleis the starting point for the personal version of this course map.

AI Course MapLearning RouteAgent Getting StartedTacit knowledge extractionSemantic EngineeringOne-person companyKnowledge Management