This article combines two things: the 'Information Time Lag Rhythm Chart' looking backward and the 'Market Prediction Three-Step Method' looking forward. Looking backward helps you know where you are now, and looking forward helps you know what will appear next month. Together, they give you a practical rhythm: prepare for next month this month, and when it appears, you're already at the finish line waiting.
- People who need to grasp the timing of topics
- People who find the market already crowded when a new feature is released
- People who push new methods within a company and are often told 'It's too early'
- People who feel they're always half a step behind and are anxious
- A six-cell information diffusion rhythm chart to locate yourself and your customers
- Monetization methods for each cell to avoid selling the wrong content to the wrong audience
- Market Prediction Three-Step Method: Find Pain Points, Promote Technology, Estimate Timeline
- A pain point reverse-engineering prompt that can be pasted into AI
Two questions, which are actually the same
People who make content or products often ask two questions.
The first one looking backward: This thing people still don't understand, will I be too early if I talk about it now?
The second one looking forward: What's the next wave, and which topic should I bet on?
These two questions look different, but their answers are in the same place: Information diffusion has a rhythm. Understanding that rhythm helps you see both where you are and where to go next.
Let me first talk about the latter half
How long does it take for something to travel from abroad to you
From my own observation, the diffusion sequence is roughly like this
- The most advanced people abroad First publish a new architecture or new approach
- About one month later Abroad will have people develop it into a practical application
- Another about one month Taiwan's more capable people will develop it into something relevant to their field. I am roughly in this stage, lagging behind the front by three to six months. After the architecture is established, I can translate it into a knowledge management version within one month
- Another about half a year More advanced companies will start to hear about it
- Another about half a year General office workers will talk about this in the office
- Another about one year The government will start promoting related courses
Converted into the lagging distance, the average is: bosses lag by half a year, general staff lag by one year, and the government's actual implementation lags by two years
The logic of the government's stage is simple: this year sees that this thing can be done, starts writing the plan, and next year executes. So when the general public is already talking about it, the government is still writing the plan. This is a structural issue with the budget system
What I'm talking about now is what I learned two years ago.
The most direct evidence for this table is my own teaching content.
Now, when I give talks or teach in government departments, the content is mostly what I learned in 2024. What I learned in 2025 is what I'm currently sharing in the community.
When I first realized this, it felt a bit strange, but after thinking it through, it became a very useful scheduling tool: What I'm learning today is what I'll be teaching in two years; what I'm teaching today is what I learned two years ago.
Therefore, I don't need to worry about 'people don't understand this yet,' because that's just the time lag. I also don't need to rush to include the latest content in my courses, because the people who need it haven't arrived yet.
Standing in different positions, you earn different kinds of money.
Once you know your position, the next step is to know what each position can do. This is the most practical part of the entire table.
There is no market here, only raw material. What is suitable here is to translate it into your own field's version, and first establish works and cases. Selling courses here is usually not successful.
Companies start to have whispers, and people start to talk about it. This is when demand is just beginning to emerge. The best-selling positions for courses, consulting, and mentoring.
Government opens courses, companies allocate budgets. The volume is the largest, and the price is stable, but the requirement is complete teaching materials and deliverable specifications, not the latest viewpoints.
A common mistake is to use the content from the first position to sell to customers in the last position, and then feel that the customers are not keeping up. In reality, it's not that they are not keeping up; it's just that they haven't arrived yet.
The opposite mistake also exists: staying in the last position, talking about content from three years ago, and then being caught up by competitors, because those contents are something anyone can talk about.
My own approach is to stand in two grids at the same time: Learning stands in front, delivery stands in the middle. Learning always moves forward, grabbing what the people abroad are talking about and academic papers. Delivery targets the grid where the demand has just started to grow, because that's where people are willing to pay. The time gap in the middle is your preparation period, which is also your moat.
Switching direction: What's next?
After knowing where you are standing, the next question is: What's next?
I can roughly predict what AI big companies will release in the next two to three months. This is actually not difficult, because big companies release features in response to demand
The logic is simple: Look at what the main demand is currently, and AI will go to solve that problem. You can roughly understand the principle of AI, and then you will know what technology they will use and what products they will release to solve it.
The same logic applies to things that will go viral. A thing goes viral because it uses a very ingenious mechanism to solve the pain points that most people are currently facing. So instead of studying how to operate this tool, you should study who's pain it solves.
From the process of studying the pain points of the masses, you will discover the trend. The trend is in the pain points of people
Three steps of prediction
The thing that people have already silently accepted and are working around with makeshift methods is more valuable the longer people endure.
To solve this pain, the model needs to do one more step. If the existing technology is just a little bit away, it will appear quickly.
Grab one to three months. The key is to design what you want to do with it before it appears. Do not guess the exact date.
The third step is where the entire method truly delivers value. Prediction itself has no meaning; it is only meaningful when there is action after the prediction.
My rhythm is like this: This month, I prepare for what will be launched next month. When I am ready, AI just happens to launch; it launches, and just when it can start being used, I have time to prepare for the next month's launch. This way, you are always a step ahead of the market, and you will never be in a rush.
A real case that actually happened
In May this year, while preparing for the company, I said that the pattern of meeting minutes would be changing.
Step one, where is the pain point. Our current approach is to record conversations and then make meeting minutes afterwards. However, I make detailed risk plans and predictions during meetings. Often, after the meeting, when the minutes are compiled, I find that a risk was not discussed during the meeting, and I have to schedule another meeting. This pain has been endured for a long time, long enough that people have come to accept it as normal.
Step two, what is needed to solve this. What is needed is real-time, not post-hoc. That is, the system should understand the meaning during the conversation and remind the user in real-time. At the time, this technology was already very close to being achieved, only lacking integration.
Step three, estimate the timeline. I estimated that real-time conversation would appear in July or August. Let me give an example of what I imagined: If AI could remind me during a conversation, 'He said he would handle a case worth ten million, but his company's capital is only two million. Do you need to confirm with him?' The practicality would be completely different.
GPT Live actually appeared in July.
My current judgment is that the second generation will appear in September, and people will start trying to integrate it into meetings and work. Because the current generation is basically just for chatting. I have already started developing a new meeting guide model with my partner who does corporate consulting, introducing AI advisors in real-time during meetings, which will also require adjusting the way organizations hold meetings. I have asked my partner to prepare for this, so that when it is launched in September, we can directly hold a workshop.
What happens if you guess wrong
This is why many people are hesitant to use this method, so it is important to explain clearly.
The betting approach of this method is safe because you are betting on whether a certain pain point will be solved rather than whether a certain feature will appear. The pain point is real, so it will eventually be solved. The difference is who solves it and when.
Therefore, the preparation you do in advance is usually one of these: think through the process, write the materials, find the partners, and collect the data you need. These things will not be rendered useless if the feature appears two months later.
What will truly be wasted is to learn the specific operation details of a particular tool in advance. That is definitely outdated, so I do not learn the operation in advance. Instead, I think about the process and application in advance.
Overlaying two timelines
The schedule and prediction method are useful when viewed separately, but when used together, they create a clear work rhythm.
Research the pain points of the general public, and calculate what will be solved next. This month, focus on designing the application. This determines what you 'learn'.
See where your target audience is in the grid, and deliver in the language and form that they understand. This determines what you 'sell to whom'.
Problems arise when you mix the two approaches: using what you learned from the front to explain to the last group of customers, and then feeling the market is not mature. The market is not immature; you have overlapped the time axis.
Three questions you can ask yourself now:
- What I am currently talking about, when did I learn it? If the answer is this year, you may be too far ahead. First, clarify who you are selling to.
- Where is my customer in the grid? Enterprises, general employees, and public departments have very different time points. The same content needs to be explained differently.
- Who will I deliver what I am learning to in two years? First clarify the target, and your notes today will have structure.
Directly copyable pain point reverse prompts
Fill in your industry and paste it to AI:
Conclusion
Most people's rhythm is: new features come out, see everyone discussing, start learning, and after learning, find the market is already full.
Change the rhythm: first look at the pain points and calculate when they will be solved, design the application before that; at the same time, clearly understand where your customers are and deliver in the way that they can understand.
This way, you won't be too late, nor too early with no one understanding.
Finally, add one sentence: the best part of this is that it doesn't need insider information. The pain points are there, everyone can see them, but most people are used to it, so they don't realize it's a problem.
Further Reading
- Solving AI Knowledge Anxiety: Finding Anchors That Remain Unchanged in a Rapidly Changing Era After grasping the anchor point, this article's rhythm table will show the direction to grab the topic.
- Why Your AI Has to Start Over Every Time: Four Steps to Teaching AI Predicted things need to be implemented, relying on the rule base, not tool operation.
- No AI transformation, only AI-native After a new feature appears, you need to decide whether to use it to optimize the old process or redesign.