AI writes articles with more than 100,000 views, relying on these secrets

In the digital age, AI technology is gradually penetrating into the field of content creation, helping creators produce high-quality articles at an unprecedented speed and scale. This article explores in depth how AI helps creators create popular articles and reveals several key secrets behind AI creation.

Today, in just over a year, the development of AI has exceeded people’s imagination. When it comes to using AI tools to create content, people are divided into two camps: one uses AI to add icing on the cake, and the other always feels that AI is not satisfactory after using it.

Is AI really not developed enough? In fact, when interacting with AI, "simply saying it is simple and making it clear are two completely different things." Many people who think it is unsatisfactory often stay in simple conversations, and it takes a long way to go from questioning to understanding.

So what capabilities are needed to use AI to create a hit article? On this topic, Jianshi invited Wang Weinan, co-founder of Xumei Intelligence, to the live broadcast room to discuss how to create a hit article and how AI can assist.

This AI practitioner, who has written more than 100,000 articles using AI, shared his understanding of content creation: first grasp human nature and then use AI thinking to communicate with AI. As long as you persist, there will be possibilities.

Next, let’s go back to the scene and listen to Wang Weinan’s understanding and thoughts. As follows, enjoy:

01 To create a popular public account, we should consider the mechanism and human nature.

The first step to creating a popular public account is to know what popular articles look like. Wang Weinan classifies "popular content" into several types.

1) Information content. Works like Daolang’s are hot information and breaking news. In the WeChat ecosystem, live broadcast accounts and video accounts are also hot information. The key is to publish quickly.

2) Emotional content. Daolang's content also contains emotional elements. As an old singer, he shows the flow of emotions. Some of the content that Ji Ge has created, including some of my own creations, are all designed to capture everyone and provide emotional value. Such emotional value can either make you rich or resonate with you and bring back memories.

3) Content that provides valuable information. This means providing some insights and showing deep thinking. After reading the article, users may have a deeper connection and cooperation with the author. This is a kind of soul communication, which is similar to the experience of reading books.

The second step is to understand what readers need and obtain referrals from the official account.

Wang Weinan summarized his own rules based on his own experience. If you want to break through the recommendation mechanism and gain traffic, you need to break through five major levels.

The first level is authority. This authority is very valuable. If you want to write a similar content today, you must first consider your own authority. If not, your chances of success will be greatly reduced under the same information background. The essence of the PK algorithm is a horse race to see who will win in the end.

The second level is high-quality fans. If you have accumulated a large number of fans with deep reading ability, you can gain an advantage in the competition for traffic.

The third level is news sensitivity. After producing a large amount of content, the most likely to cause explosive news is breaking news. If you get information and create relevant content in the first place, you will have an advantage.

The fourth level: how much content should be written. The fifth level is whether to use AI system to write. The purpose of using AI is nothing more than to improve efficiency. Originally, when writing courses or content, you may only be able to complete 5,000 words per unit time, but now with the help of AI, you can reach 20,000 words.

So how can AI assist in this process?

One of them is to build a super large content library. Use technical means to organize the content library, including references, theoretical basis, related books and learning content. Especially in data integration, AI is much more efficient than manual work.

The second point is to help editors improve content and topic selection through understanding of platform mechanisms and analysis of big data. In the future, AI will have more efforts in this direction.

The third point is that after manually screening the styles that users like, AI helps creators batch generate content that users like within a given style model, including interactions with users.

This brings us to the third step in creating popular public account articles: resonating with users.

Wang Weinan mentioned that whether it is AI-generated content or manually written content, the most important thing is whether it can resonate with and interest users. The reason why people like their own content is a two-way choice. If someone is not interested in technology or data content, they will naturally not pay attention.

In addition, your own content system, whether or not it is generated by AI, must maintain a consistent style. Even if the content is generated by AI, most friends will find it difficult to distinguish.

02 Use AI thinking to communicate with AI

After considering human nature and some necessary conditions, how AI can empower our articles becomes the key point.

Wang Weinan said that in fact, as long as the appropriate instructions are set, it is relatively easy for AI to generate content. However, why can't most people do this? Because many people lack communication skills, which makes things complicated. When many people want to make money, they simply give the big model an order: "I want to make money." In doing so, they lose most of their users, and there is no effective way to change this situation.

How to communicate with AI? Wang Weinan put forward two important viewpoints: code thinking and reverse thinking.

The concept of code thinking is proposed to adapt to multi-round dialogues with large models. In the process of repeated communication with AI, high-quality content can be produced.

Code thinking can be understood as the existence of interrelated relationships between paragraphs A, B, and C in the article. These relationships may be progressive, parallel, or inclusive, or even other complex relationships. In these relationships, there may be multiple ways of connecting paragraphs and sentences.

The key is to understand the relationship between A, B, and C, and how they interact. Whether it is a word, a paragraph, or a sentence, it needs to be constantly defined.

For example, when a Jianshi article mentions Daolang, what role does the data it displays play in the entire article? How is it related to the current hot topics?

In the next article, we may touch upon a similar topic about another singer. Daolang is essentially a singer, and this "singer X" can be Daolang, Jay Chou, or other singers such as JJ Lin. These choices also easily reveal age differences.

So how much weight does the data have on singer X? For example, a singer is live streaming on a video account. Then, his data will interact with Daolang. If I regard Daolang as X and Jay Chou as Y, then when Jay Chou is live streaming on a video account, his data compared with Daolang's data will show the difference between X and Y.

By comparing these data models, a new article is actually formed. The key to code thinking is to ensure that the content is not made up at will, but always has a theoretical basis.

The other is reverse engineering thinking, which is to conduct reverse analysis and research on the target account, so as to deduce and derive the account's processing flow, organizational structure, functional characteristics and other design elements, in order to create products with similar functions but not exactly the same.

It is not a simple imitation or plagiarism, but a way to deeply understand the composition of things. When you want to create something, the most effective way is to analyze its components first. The core of reverse engineering is actually understanding and transcendence. Through this process of disassembly, the key steps are three points.

First, you need to have a clear idea. For example, through discussion and trial, you can determine which steps are feasible and which ones need to be adjusted.

The next step is to determine the goal and collect information. Analyze the copywriting and video content of the big bloggers, and understand the classification of different content, such as the news system, the dry goods sharing system, the post-interview manuscript system, and the live broadcast sharing system. At the same time, test whether each content can attract the audience.

The third step is to consider designing some of your own content to make your own things richer and more suitable for the reading habits of your users.

03 Big pitfalls you must beware of

Although AI is an inevitable trend in the future of content creation, the development of history is always a spiral curve forward, and even AI will have many unavoidable problems. Wang Weinan said that if you want to get involved in AI content creation, you need to avoid the following pitfalls.

The first pitfall is to avoid being blind, to know how to find clear trends and stick to them.

The key to learning is practice, because it is extremely valuable. Sometimes we put a lot of energy into exploring and understanding, but if we don’t put it into practice, we can’t really solve the problem. Many people always look to successful people or wealth accumulation for some vague answers, but this is not the case.

Take myself as an example. The achievement of more than 100,000 posts actually required a lot of continuous persistence and exploration. AI can amplify the potential of editors, but it cannot omit the experience of editors’ personal practice. If you think that you can avoid hardship with AI, this idea is not advisable.

At the same time, choosing the right era and trend is also very important. A clear trend is crucial. You must recognize this trend and stick to it to see results. Take the AI ​​track as an example. Many peers use AI technology to chase hot spots in the payment field. Although it attracts attention in the short term, it ultimately fails to last because they do not see the big trend of AI.

They just thought they could make money in this area, muddle through, and eventually exit the market. The same is true for content and AI. If you just want to make quick money in the short term, it is easy to look for new directions after a brief taste, which will accelerate your fate of being eliminated.

The second pitfall is about small advertisements. Many people will tell you that you can get everything for 99 yuan, but in fact, this is misleading. Opportunities exist, but there is a big probability problem.

Is the content valuable? It depends on yourself. For example, using AI to provide in-depth content services for B2B companies usually has great commercial value. If AI is used to create emotional content, such as a psychological counselor, this may also be valuable. But many people always package any content as something valuable. Many people who write books and content often emphasize that they are counterattackers in a fourth-tier city, but two days later they say they are old actors.

This kind of statement is a kind of "cutting leeks" content, especially in the context of artificial intelligence, the problem is more prominent. Therefore, we should choose those contents that have long-term stability and high technical content, even if they sound boring, they are worth learning, rather than those that claim that you can solve the problem for 99 yuan.

The third pitfall is that everyone must make some practical attempts. Otherwise, there is no way to move forward. Many people ask how to solve problems every day, but when they get the answers, they don’t believe them and still want to do it according to their own theories.

Wang Weinan also mentioned that among the three major "pitfalls" he observed, many people used large models to create content at the beginning, but did not persist. In fact, in the next one or two years, AI will gradually take control of this field, and if you don't persist, you are likely to be eliminated.