If we go back to 2020, when ChatGPT first came out, few people could imagine that by the beginning of 2026, there would be an offline door-to-door service: helping people install and deploy AI Agents.
After the emergence of the Internet, we have experienced many trends: from e-commerce, self-media, cryptocurrency, all the way to today's AI. At first glance, AI seems to be just one of many new things, but in my opinion, it is fundamentally different from the previous waves.
E-commerce addresses the need for shopping, self-media addresses the need for information dissemination and expression, and cryptocurrency addresses the need for financial decentralization. These technologies have truly changed our lifestyles, but they have not directly changed "us as people". What AI wants to solve is our need for “knowledge”.
In this era, it seems that we no longer need to study step by step for many years before we have a chance to meet a good teacher. A question-and-answer ChatBot is like a knowledgeable tutor who knows almost everything; an AI drawing tool that can generate beautiful paintings through natural language; Wensheng videos that are difficult to distinguish between authentic and fake (such as Seeddance 2.0); and a complete set of code capabilities that can generate websites and applications in just one sentence - these are all directly rewriting the way we acquire and apply knowledge.
This means that the past path of “spent twenty years of hard study in exchange for a little knowledge advantage in a certain field” is being rewritten. We used to rely on the scarcity of knowledge to make money: think back to those high-income professions in the past - doctors, lawyers, programmers, all positions that required extremely high professional knowledge. However, now, with the advent of AI, it seems that ordinary people can also "borrow" these abilities to some extent: the technical threshold for programmers has been lowered to "making a website in one sentence"; lawyers can first have an in-depth conversation with AI for professional consultation.
So, does this mean we don’t need to study anymore? In times like these, what is truly important?
This is exactly what I want to discuss next.
In my opinion, time, decision-making and growth are the three most important things in this era.
Time
In the era without AI, the biggest gap between people was the knowledge gap brought about by education. But the traditional education system is inefficient: much of what you learn in college is often far from the abilities you need in real work. Ultimately, you still have to relearn at work. Academic qualifications are more of just a "stepping stone" to prove that you have the ability to learn and adapt to new knowledge.
But things are completely different now. Imagine two primary school students who started first grade at the same time: one comes home from school every day with his iPad in his hand and watches short videos; the other, under the guidance of his parents, uses AI to assist learning, explore the world, and develop his own interests and hobbies, which is equivalent to having a patient and top-notch private teacher by his side all the time. Ten years later, when they grow up, how big will the gap between these two people be? Not only will AI not smooth out the differences between people, but it will infinitely magnify the differences due to the great improvement in efficiency per unit time. The difference in the future between those who continue to use AI proficiently and those who only watch short videos is likely to be like the difference between Homo sapiens and monkeys.
Therefore, in this era, how to "spend time" has become particularly critical: are you using time to "accumulate volume" for yourself?
Many people are aware of this change, and what follows is anxiety, which is the so-called FOMO (Fear of Missing Out). The "crayfish" thing has become very popular recently, and many people have bought Mac mini to build their own "crayfish". But the reality is that most people don’t even know how to use the command line, but they imagine that this thing can immediately improve their productivity. The Internet is full of similar content: Installing a certain skill can increase efficiency by 100 times, or installing a quantitative trading skill can make $400,000 a week. It seems that as long as you deploy the "crayfish" well and copy other people's skills, you can become rich overnight.
The extremely high number of collections of this type of content just reflects the current general impetuosity and anxiety: it seems that if you don’t install this “crayfish” today or quickly follow a certain skill, you will be left behind by the entire era tomorrow. But they completely didn’t realize that our overall efficiency is far from the level of “doubling it 100 times with just a few clicks”.
Most people's understanding of time still remains at the level of "a line": from birth to death, it is one-dimensional, like a ray. So when facing new things, their thinking is: as long as I participate once in this timeline and do one thing at a certain node, I should reap rewards. So, I spent 500 yuan to hire someone to install "Crayfish" at my door, and then opened a certain quantitative trading skill in my favorites. I didn't even know how to install it, so I paid someone to help install it, and then hurriedly invested some money in it. After losing all, I cursed AI as a "scam." This is a typical one-dimensional biological perspective: it only focuses on the "operation" of a certain moment on the timeline, rather than the "investment structure" of the entire time dimension.
But time is closer to a "pipe", which has a cross-sectional area and a volume. Think back to your life experience: Was there a period of time when you were particularly hard-working (such as your senior year in high school, when you were rushing to take the college entrance examination), or a period when you were particularly wasted (such as your relatively relaxed college years)? This is the difference in the "thickness" of time: during high school, the cross-section of the pipes was thicker; during college, the pipes were much thinner. Then multiply the "cross-sectional area" by the length of time (three years in high school, four years in college), and you will find that the "volume of time" at different stages is actually completely different.
Look at those behaviors from this perspective: You want to spend a day and 500 yuan to hire someone to install a "crayfish" remotely, and then randomly install a skill you don't understand at all, just to change the entire trajectory of your life. This thing itself seems very ridiculous.
For a long time in the past, the efficiency gap between people in "unit time" was not large, that is to say, the difference in the cross-sectional area of the time pipeline was limited. More differences come from the length of time and continuous investment, so many people feel that "everyone will be the same in the end", so there are jokes like "the world is a big grassroots team".
But after the emergence of AI, the gap in unit time has been widened to the level of "heaven and earth": On one side, people are watching short videos with their mobile phones, and the cross section of the time pipeline can almost be regarded as a point; on the other side, they are using the same hour to interact with the world's top knowledge and tools at a high density.
What really matters is not whether you followed the trend and installed "crayfish" today, but whether you use AI to thicken and lengthen your time pipe, or flatten it into "time fragments" chopped up by short videos.
Decision-making
The second thing that I think is equally critical is "decision-making".
Whether it is the current social situation or the speed of technological evolution, we are already in a highly turbulent and rapidly changing era: new events happen every day, and new technologies emerge every day. In such an environment, how to identify the direction and make relatively correct choices has become the core ability that affects the direction of life. It can be said that the quality of decision-making determines the quality of life to a large extent.
To give you an example: A colleague and her husband casually made a "plan" at the end of last year - give it a try at Christmas, and if they get pregnant, they will let nature take its course and have a second child. If not, they will continue to have one child. Regardless of whether there is scientific preparation for pregnancy or not, such a major event that can completely change the trajectory of life is actually finalized with an almost bet-like mentality. As expected, she got pregnant. Nowadays, she suffers from morning sickness and anxiety every day. While working, she laments: The money she makes now is simply not enough to raise two children. This is the epitome of the quality of decision-making, which directly feeds back to the quality of destiny.
We think about it every day: big things, such as where the world will go; small things, such as what to eat at night. After thinking is over, decision-making comes next; decision-making brings action, and actions accumulate little by little into our destiny. Therefore, how to make better decisions is by no means a "chicken soup problem", but a very real survival problem.
Here I recommend a book: "The Art of Decision-Making". This book provides a relatively complete and operable decision-making framework. It is not a theory that will be left on the bookshelf to collect dust after reading, but a process that can be used immediately. After reading it, choose something you are struggling with, follow the steps in the book to make a complete decision-making process, and practice again and again. You will find that the probability of making wrong decisions will be significantly reduced. At least, you are "making a choice after weighing" rather than relying entirely on your current emotions and intuition.
So in this era of AI, what more can we do?
We actually have an unprecedented decision-making partner - "the most diligent and tireless think tank in the world." It can help you organize information, list options, estimate risks, and give suggestions. If you extract the decision-making framework in "The Art of Decision-Making" and feed it to AI, and let it help you think about problems according to this structure, you can build your own "decision-making workflow": every important decision must be made strictly according to this process.
But the premise is that you must first understand this book and this framework yourself - otherwise you will not even be able to ask "questions to the AI". AI can greatly improve the efficiency of the "output end", but the understanding and judgment of the "input end" are still in the hands of humans.
After making the decision, the matter is actually not over yet. What really widens the gap is the link of "review + refining principle".
For example, when I bought a car in 2020, I did not buy full insurance out of the mentality of "save what you can". At the end of 2024, the car had a serious accident and was basically declared scrapped. After the accident, I made a series of the same ill-considered decisions: I wanted to save money, so I found a cheap mechanic and spent 3,000 Australian dollars on repairs. As a result, I discovered a week later that there was a major problem with the engine, which required at least 5,000 Australian dollars to replace. In the end, the car could only be disposed of according to its salvage value, and was recycled for 2,000 Australian dollars. From accident to scrapping, it was the result of a series of bad decisions.
In retrospect, there are at least a few layers of problems:
- In order to save full insurance premiums at the beginning, I exposed myself to a high-risk range. On the surface, I was "lucky" to have been driving for 5 years before getting into an accident. Looking back at the result, it seemed that it didn't matter whether I bought full insurance or not. But if the accident occurs in the first year, the value of the entire car is instantly reduced to zero. At that time, I was only focused on how much money I could save every year, and I didn’t realize that I was taking on “tail risk”—which would be a fatal blow if it happened.
- After the accident, I still use the same bad mentality to make choices: I continue to be "cheap" and look for the cheapest car repair solution, but do not carefully evaluate before making a decision: If I find a bigger problem halfway through the repair, can I still afford the higher repair cost? Is this car worth continuing to throw money into? A more rational approach might be to stop losses early and calculate all insurance or subsidies, residual value and future risks together.
- The whole process has almost no plans and boundary conditions: under what circumstances must we admit losses and stop losses? Under what circumstances is it worth continuing to invest? All of them are based on feelings on the spot.
This is the price of "not reviewing decisions": the same mistake will be repeated again and again in different scenarios.
Therefore, after making a key decision, the most important thing is to sit down and review it, extract a few principles that can be written into your own "life operation manual", and record them. The next time you encounter a similar situation, you will no longer just "just let it happen", but will have a gradually polished decision-making system of your own supporting you.
AI can help us collect information, build a framework, and even remind us to review, but the person who is really responsible for decision-making is always the person on the other side of the screen.
Growth
Since AI is already so powerful, do we still need to study seriously on our own?
If you read the previous two parts of "time" and "decision-making" carefully, the answer is already obvious: my thinking is not just based on talking to AI. I still continue to invest a lot of energy in reading and use systematic knowledge to upgrade my underlying abilities. Only when your own cognitive level is continuously raised, can you ask better questions and issue clearer and more constrained instructions, can AI have the opportunity to give truly valuable answers.
My understanding of "time" comes from "The Truth about Wealth"; the decision-making framework I use comes from "The Art of Decision-Making". In other words, there should be high-quality input first, and then high-quality interaction with AI, rather than leaving everything directly to AI.
Here I want to introduce a concept: Reading efficiency.
can be roughly understood as: how much knowledge increment and cognitive improvement you get from text in the same time. If they spend the same 30 minutes, some people are reading short and quick information fragments, and some people are reading a chapter of a good book. In the end, the "reading output" of the two people is completely different. Behind this is the difference in reading efficiency.
Books, especially good books, are often hundreds of thousands of words condensed by the author using several or even ten years of experience, research and reflection. You can "download" what others have seen and thought for years in just a few hours, which in itself is a very high level of leverage. A Weibo or a tweet is often just a sentence of feelings written by the author in a few minutes. If you spend the same reading time on both, the structured cognition and systematic perspectives you can obtain are obviously not on the same scale at all.
Therefore, from the perspective of "return rate per unit time", the reading efficiency brought by system reading is almost certainly much higher than that of fragmented content.
This is why, in the AI era, we need to read more. Reading is not to compete with AI for "who knows better", but to empower yourself to use AI to a higher level. The more you understand, the more specific your questions become, the clearer the constraints, the more realistic the scenarios, and the more in-depth and realistic the AI’s answers become. And if your own understanding remains at the surface level, and your question can only be a vague sentence such as "I have a problem, please help me solve it", then the AI's answer will most likely only be general suggestions.
You can make a simple comparison:
One person has read "The Art of Decision-Making", thoroughly understood the framework inside, organized it into clear steps, and then sent it to the AI, so that it can assist him in making decisions strictly according to this framework;
The other person has not read anything and will only say to the AI: "I have something, please help me analyze how to choose."
Both people are using AI, but the quality of help they can get is almost completely different.
To me, AI is more like a mirror and an amplifier: wherever your level is, it will amplify that level exponentially; the more solid your foundation, the further it can help you go. Learning is never because AI is not strong enough, but because if you don’t grow, you will never be able to use the true upper limit of AI.
When it comes to "growth", it ultimately comes back to "time".
Time is our most fundamental means of production. We use time to learn knowledge, and then use knowledge and tools to ask for it from nature and the world, meet the needs of others, and complete production; the goods or services produced are eventually exchanged for money and opportunities. All these processes are essentially a "reprocessing of time."
So, please remember: time is not a one-dimensional line, but a "pipe" that has both "section" and "length". The so-called growth means that while improving efficiency, the "cross-sectional area" of the pipeline becomes larger; while maintaining iteration, the pipeline continues to extend in the time dimension.
It is more intuitive to use two words to explain this kind of iteration: iteration, or even more accurately, "recursion".
In computers, recursion means: taking the last output as the next input, over and over again, layer by layer. The same is true in life - every time we practice deliberately and every adjustment after review, we are providing better "input" for the next action.
For example, running every morning is the simplest recursion:
On the first day, your physical condition is the "input". After a running training, you get an "output"; on the second day, you continue running on this basis, treating the previous day's output as a new input; day after day, your physical fitness, endurance and mental state are constantly superimposed and amplified in this recursion, and finally show a very intuitive growth.
The same goes for learning. Read a book, take notes, have an in-depth chat with AI, and write a summary. This whole process is a recursive unit. You can choose to make time a "line" that passes passively, or you can make time a "channel" that continuously strengthens itself. The only difference is whether you realize that your daily "input" and "output" are forming a positive recursion.
When you really start to use time, use AI, and use reading like this, you will find that the so-called growth is no longer an abstract big word, but something that can be seen, calculated, and felt every day.
Summary
I am probably one of the earliest paid users of ChatGPT. From the beginning of the AI wave to the present, it took me several years to slowly figure out: what we really need in this era.
This is an era that has never appeared in human history - Everyone has the opportunity to expand their abilities, and everyone can acquire knowledge with almost no threshold. Those resources that once belonged to only a few people are now lying there quietly: models, tools, courses, books, cases, and role models, all at your fingertips.
The difference has never been about "whether there are opportunities", but rather: some people take the initiative to reach out and use time and action to turn them into their own abilities; while some people choose to turn a blind eye and let the dividends of this era blow past them like a gust of wind.
AI puts the matter of "opportunity" in front of almost everyone for the first time. The next gap in your life depends on how you are willing to use your time, how you make specific choices one after another, and whether you are determined to let yourself grow continuously and recursively.