In den nächsten 10 Jahren gibt es 4 große Chancen für Durchschnittsbürger, ihr Schicksal zu verändern.

In the era of AI, some people are anxious about being replaced, while others see new opportunities.

The essence of AI is not to replace humans, but to make "human experience replicable", thereby giving rise to complex services that were previously impossible to scale. From health management and transportation to personalized consumption, a wealth - creating wave centered around "service scaling" is unfolding.

In the era of artificial intelligence, the real winners are not those who understand technology best, but those who understand needs best and can use technology to meet needs on a large scale. Just like Liu Bang, one should have the ability to predict and coordinate resources, and use AI (like Zhang Liang and Han Xin) as a powerful tool to jointly provide large - scale complex services for users, thus taking the leading position in this wealth - creating wave.

This article will break down the key logic of combining AI with industries to help you find your own ticket to enter the game.

It is hoped that it can inspire you.

I. The essence of AI is the "scaling of human experience"

The core opportunity in the AI era is to make human experience replicable and give birth to a brand - new business model of complex intelligent services.

1. Two stages of AI development: from energy - saving and efficiency - increasing to market expansion

Actually, current artificial intelligence can already accomplish many tasks, but its main function is to save energy and increase efficiency, that is, it is cost - saving.

You know, cost - saving businesses can lead to GDP shrinkage. Because for something that originally cost 10 yuan, now it can be done with only 5 yuan, resulting in a GDP loss of 5 yuan.

Therefore, I think Cathie Wood was ultimately too optimistic. It's not that artificial intelligence is useless. Even for a useful technology, its impact on GDP can be divided into two types: one is cost - saving, and the other is market - expanding. Only when market - expanding applications are widely adopted can GDP truly grow.

So, the real greatest advantage of technology lies in market expansion. Cost - saving businesses are essentially a replacement of existing businesses, and existing businesses often don't perform well.

2. Historical laws confirm: all great technologies will eventually create new markets

First, Watt's steam engine created the markets for trains and steamships.

The original intention of inventing Watt's steam engine was to replace Newcomen's steam engine. However, only about two or three hundred units of Watt's steam engine were sold during its life cycle, mainly because not all regions had a large number of mines.

Even in the UK, a country rich in coal resources, compared with the European continent, the number of mines was relatively limited. Therefore, you'll find that, frankly speaking, the demand didn't explode on a large scale in that era.

So, when will the demand increase significantly? Obviously, the explosion of demand doesn't just come from replacing existing equipment, but from new demand. Specifically, the emergence of trains and steamships was the key.

Watt registered the steam engine patent in 1769, but the experiments on trains basically started around 1800. By 1810, commercial B2B freight trains appeared, and in 1820, B2C passenger trains arrived. The appearance of steamships was also around the same time and took about another 40 years to be widely used.

This is the emergence of new demand and new applications to meet new market needs. Subsequently, the demand increased significantly because the demand for trains and steamships was far more than 200 units, but a huge global demand, so the market expanded rapidly.

Second, Ford's Model T created the mass - market for automobiles.

Similarly, take automobiles as an example. Although Benz is regarded as the manufacturer of the world's first automobile, its significance is not significant because it couldn't achieve mass production and could only replace carriages to a limited extent. Suppose there were 1000 carriages used by German nobles, Benz might only replace 100, 200, or 300 of them.

However, Ford's approach was completely different. Ford didn't just replace, but created new market demand. For those who originally couldn't afford or use carriages, Ford cars were extremely affordable. Calculated at that time's prices, Ford workers' daily wage was 5 US dollars, and the cheapest Ford car only cost 360 US dollars, which meant that one could buy a car with 72 days' work income.

It's worth noting that this refers to ordinary workers on the production line, not the middle - class or executives, and the price was amazingly low.

Therefore, when people saved enough money, Ford cars could be seen everywhere on the streets. In a market where there were originally no cars, Ford successfully developed this emerging market.

As we all know, the development of emerging markets is the most difficult, and market demand is difficult to form a scale quickly because people don't have obvious needs at first. However, from 1908 to 1927, within 20 years, 16.5 million Ford cars were sold. This achievement is remarkable, and this is what we call the "newly - added" market.

3. The essence of the AI revolution: service scaling

The most fascinating thing about the industrial revolution is that it always starts with replacement. However, this also triggers strong negative voices. Why? Because replacement often leads to the disappearance of the original things and the unemployment of employees, which makes people have complex feelings.

But in fact, this is not the essence of the technological revolution. The real essence lies in the subsequent transformation, that is, from energy - saving and efficiency - increasing to the creation of new markets and new applications. Only at this stage can we see the truly remarkable aspect of the technological revolution, and this remarkable aspect is explosive.

In fact, when we look back at technological revolutions, many patterns are clear. We often repeatedly talk about "the rhymes of history", which means that history really repeats itself, not in terms of content, but in rhythm, just like a rhyme. Each technological revolution relies on a powerful core technology.

In the past, this core technology was the steam engine, that is, mechanical power. Simply put, human power could never reach the level of mechanical power, which was a leap. From then on, trans - oceanic voyages and travels became possible. So, the world suddenly became smaller, but the demand suddenly became larger.

Actually, each technological revolution has such a characteristic: its real highlight lies not in energy - saving and efficiency - increasing, but in the generation of new demand. The new demands brought by each technological innovation are different. This is just a similarity in form, not a simple repetition.

So, what is the essence of this round of technological revolution? In the previous round, we said it was mechanical power, and in this round, the core is artificial intelligence. The essence of artificial intelligence is that human experience can be replicated.

Previously, mechanical power combined with assembly - line production achieved the replicability of products, which we called production scaling. In this round, it is the replicability of human experience, which we call service scaling. Through replicating experience, we have entered the era of service scaling.

II. Future scenario: characteristics and implementation paths of complex intelligent services

1. Four characteristics define the next - generation AI services

Services have always been a big problem: high - end services are out of reach for most people because high - end experts are scarce, the market is niche, and the prices are expensive; while general services are of low quality and cannot meet complex needs.

Therefore, it is difficult for services to be large - scale, complex, and personalized at the same time. However, artificial intelligence is different. It is a matter of computing power.

AI can provide expert - level, personalized, continuous, and inclusive services and achieve scaling. Next, we will focus on continuous services, expert - level services, and inclusive services.

① Continuous services

What is continuity? It is not a one - time event. For example, even the most experienced traditional Chinese doctors can hardly accurately diagnose the root cause of a disease at a glance. Why? Because there is a lack of in - depth understanding of the patient's background.

However, if we introduce the concept of continuous services, we can even imagine that when the era of continuous services fully arrives, not only this generation but also their descendants - the second and third generations - will use the same services.

In this way, the service system will have an extremely in - depth understanding of individuals. Take health management as an example. It can even master family medical histories, and its accuracy may far exceed that of doctors.

Therefore, you'll find that this phenomenon is amazingly similar to the past, or it is a "rhyme" of history. This means that AI services will reach a new height that humans can hardly achieve.

② Expert - level services

How should we understand expert - level services? After AlphaGo defeated Lee Sedol, there was no more suspense. From then on, AlphaGo left Lee Sedol far behind. With the emergence of AlphaZero and subsequent new versions, the possibility of a comeback no longer exists.

Imagine, can Ke Jie mutate one day and defeat artificial intelligence? Don't even think about it. It's impossible. Because the evolution speed and technological progress of artificial intelligence far exceed those of humans.

This also means that the profession of Go coaches, especially top - level coaches, no longer requires human labor, and artificial intelligence can do the job. For Go learners, this is undoubtedly a blessing. In the past, we may all have hoped to be personally taught by Nie Weiping, but Nie Weiping couldn't teach so many people. Now, artificial intelligence makes it possible for everyone to receive instruction.

Now, there are numerous powerful artificial intelligence programs everywhere. AI products for teaching Go and Chinese chess are all over the place, easily accessible, and inexpensive.

This has led to another interesting phenomenon: once - scarce and expensive expert - level services are no longer scarce, so the prices naturally drop. Coupled with their low hard costs, the prices will definitely be very cheap. Thus, society is becoming better.

③ Inclusive services

What is inclusiveness? Simply put, it means the services I need are not only inexpensive but also widely available, and anyone with a need can get them.

This means that a large number of services needed by the public will be transformed into commodities and enter the market, especially those newly - added service items that didn't exist before but now do, and the demand is huge. Once they enter the market, just one item can be sold in large quantities.

Note that this model doesn't charge a high fee from one person, but a little from each person. Since the user base is large, for example, if the user group is as large as billions, even if each person pays only one yuan, the total amount will reach billions.

Moreover, what if more services are added on this basis?

Think about how many affairs in our lives require professional knowledge? From career development, learning and research, and corporate business strategies to household organization, garbage disposal, and even daily health care and education, all need the support of experts.

But, have you noticed? In all these places where expert help is needed, basically, there are no experts to help you now because human labor is expensive, and with social progress, the cost of human labor is getting higher and higher.

So, the only option is self - service, training yourself to be an expert.

2. Social scenario change: from buying products to enjoying services

When we try to train ourselves to be experts, a paradox arises: How can I be an expert in every field? So, we can only make do and become half - baked experts.

So, why do we buy products? Actually, we are buying the services behind the products. However, because large - scale services couldn't be provided in the past, the approach was to solidify services into products so that we could serve ourselves through products.

I believe everyone has had such an experience. For example, when you need to do some minor repairs at home and saw a steel pipe, you specifically bought a saw for steel pipes because it was different from ordinary saws, and ordinary saws couldn't saw steel pipes. After buying it, you used it once and achieved your goal. Since it wasn't expensive, you just put it aside and never used it again.

Have you also bought many such one - time tools? And you'll find that even if you suddenly need to use it next time, you may have forgotten how to operate it, either can't find the tool or have forgotten the usage method, so you have to buy a new one.

I guess everyone does this. What does this indicate? What you really need is not the product itself, but the service behind it.

During the industrial revolution, a core element of large - scale products was self - service.

So, a good product doesn't need a bunch of complex instructions. Users can figure it out just by touching it, and that's a good product. A product that requires a bunch of instructions and complex operations to learn is a bad product.

Now, people are gradually understanding what a good product is: what users really need is not the product, but the service. As long as users can serve themselves easily, it's a good thing.

Finally, we have entered a new era where we no longer need to provide products but can directly provide services. For example, if users need health services, what I should sell you is not a blood - pressure monitor, a blood - glucose meter, or even nutritional supplements. Why? Because these may not be what you really need.

Instead, I now know what you need and can tell you how to exercise, adjust your diet, and arrange your sleep today. If you need a certain diet, I'll provide it; if not, I won't force it. Such services will naturally satisfy users.

Now, users don't even need to understand the underlying principles because they think you are more professional than them. This is a relationship of trust. You'll find that if you don't trust someone, you'll challenge them; if you trust them, even admire them and have confidence in their credibility, you'll follow their advice because you are not professional and they are more professional than you.

In this case, the market is naturally broad. Everyone will follow medical advice and professional suggestions, and of course, the results will be the best because they are really professional.

3. Technological foundation: Why is now the starting point of the explosion?

Many of us have found that this is the real future society. Now, the future society seems to be getting closer to us. One of the key points is that the key supporting capabilities have been solved.

First, it's complex reasoning.

Just one - time interaction is not enough. Two complex interactions and logical reasoning are needed. Now we all know that logical reasoning involves a chain of thought, and the chain of thought is no longer just talk. With the so - called "deep research", it's not just a verbal description but can really provide accurate reasoning.

Therefore, the solution to complex reasoning is no longer to solve problems through one - time intuitive or one - on - one interactions, but to solve problems systematically after in - depth understanding of the whole situation.

Second, it's long - term memory.

If you forget something I told you the next day, it won't work. The current large - scale AI models also have this problem. Ideally, we hope that even what was said ten months or ten years ago can be remembered.

So, what does this long - term memory rely on? It relies on large - scale input and output, that is, millions of token inputs and outputs.

Why is that? Because unlike humans, whose mental labor often combines storage and calculation, AI is also moving in this direction, but there is still a long way to go.

So, how to achieve it? The method is to record all previous interactions and then feed them back to you, which is equivalent to quickly replaying the previous interactions every day because you can't remember. After replaying, you'll remember. In this way, large - scale input and output are used to replace storage.

Finally, it's the ability to call third - party functions, that is, the MCP protocol.

For example, after completing the analysis, if I find that the user needs to order a meal, I need to call a meal - ordering app to order the meal and also call a payment tool to pay.

Finally, with just one command from you, I can complete all the work and tell you when the meal will be delivered. You just need to wait. Obviously, this requirement can't be met by a single large - scale model independently and requires one - on - one calling ability for support. Now, this calling ability is also becoming more and more perfect.

4. The development direction of artificial intelligence: domain experts (not general AGI)

In fact, technologies such as function calling (MCP protocol) still have defects at present. That is, although multiple functions are called, each function is not general artificial intelligence (AGI), but specialized artificial intelligence. That is, a certain function is professional in meal delivery or payment.