First, a question for everyone to think about: Why must ordinary people become AI content creators in this era?
Hello everyone, I am Wang Dalang. I'm so happy! My first post a few days ago gained 1,500 followers. I want to thank everyone who liked and followed me. From now on, I will carefully read every comment and help you answer questions related to cross-border e-commerce and AI content creation. Let's communicate and progress together.
Today I went to Shanghai for WAIC. Like many of you, I didn't get a ticket, but through my AI account, I participated in something called the WAIC FutureTech OPC Independent Pioneer Challenge (why is this name so long...) and was selected for the top 100, receiving a two-day complimentary ticket. Sigh, I've lost count of how many times I've tasted success because of being an AI content creator...
The certificate and Orange Pi sent by WAIC
Before heading to Shanghai, I wanted to write a super high-value guide on being an AI content creator. The article might be a bit long, so I recommend bookmarking it. It covers my personal insights into AI content creation, as well as tips for starting accounts on domestic platforms, including account setup, content selection, video filming and editing, censorship bypass techniques, and commercial monetization.
In this article, I will explain the entire framework clearly, and I will continue to break it down in subsequent articles.
Have you noticed a phenomenon? The relationship between AI products and traffic is like fish and water. Many excellent developers create great products, but they either die because they have no traffic, or they are killed by those with traffic and capital (big companies) who copy them. In the end, products with traffic survive, and those without die one by one...
Not even mentioning foreign countries, let's talk about China. Big players like Alibaba, ByteDance, and Tencent, as well as rising AI software and hardware stars like KIMI, Minimax, Unitree, and Rokid, are all very, very, very hungry for traffic and users, spending crazily on online and offline promotions. Elevator ads for Workbuddy, Rokid glasses, and Qianwen APP... brainwashing AI ads will only increase in the future...
Under this massive capital investment, the 'Hangzhou Five Phoenixes' living in the Hangzhou Cloud Valley Center were born: Datawhale, Way to AGI, Guan Cha, SegmentFault, and TGO Kunpeng. They all made a fortune through AI media. Of course, besides these AI media organizations, under the radiation of this unprecedentedly huge AI promotion fund, even small AI content creator KOLs or even KOCs can earn several times more than other content creator tracks. An AI blogger on Xiaohongshu with 20,000 to 30,000 followers can earn 100,000 RMB per month.
Visiting the Hangzhou Cloud Valley Center; the whole building is full of Agents and robots.
In the AI era, making a good product is no longer an insurmountable gap. Only those who master traffic can make a good product.
I don't know if you've used an AI canvas tool called Tapnow. Currently, its MAU has exceeded one million, and it has launched several AI movie-level competitions with Bilibili, and is recently cooperating with AirCity. Tapnow is currently at the same level as Libtv and Lovart under Liblib, which raised $300 million and is valued at $2 billion.
But do you know? Tapnow only started at the end of 2024. I am fortunate to know Klaus, a core member of Tapnow and a Peking University undergraduate, and I participated in the use and promotion during Tapnow's early internal testing phase. To be honest, when I first used the internal test—it was bad! It was the kind of product that could be made with current Vibecoding and a little technical knowledge! Awkward UI, laggy interaction, and nodes that would get lost from time to time. The product then and the product now are worlds apart.
But Tapnow's strength lies in its deep understanding of traffic. A young but small and beautiful team allowed them to know the power of community and self-media. Tapnow initially found many domestic KOC bloggers across various industries and fields, let everyone use it for free, and gave enough advertising budget to promote it. During use, they also patiently collected feedback from KOC bloggers and iterated daily. At the same time, they issued a large number of invitation codes and built a community of 100,000+ users. By giving computing power for posting and making workflows public on Tap TV, they gathered a large group of early users. At this point, they optimized and adjusted step by step, turning a once difficult-to-use product into a good one, and then into a leading one. Anyway, with traffic and users, what's the rush?
They know deeply—in the AI era, anyone can make a good product, but once user habits are formed, competitors cannot steal them away. Now you can build a simple version of WeChat in minutes with Codex, but what about the traffic and users?
If you are an ordinary person, please become an AI content creator in 2026.
Having said all this, I believe you can connect the dots. Why do we need to do AI self-media? Because we need to seize the users and traffic of the AI era. Whether we understand technology or coding, whether our expression skills are good or bad, whether we are introverts or extroverts, and no matter what stage of life we are in. As long as you want to get on the fast train of the AI era and welcome the arrival of AGI and robots entering thousands of households, you should create your own AI self-media.
Every follower we gain is a user in the AI era, and every short video, graphic, tweet, and interaction is our asset in the AI era. In the environment of AI + various industries, we will have our own voice in our industry. And these will return to you in the short, medium, and long term as cash, connections, and resources.
To put it simply, if you are a developer who understands technology, your product will have users because you did AI self-media and gained traffic. Not only will you generate subscription income, but you will also receive AI review ads from large and small factories, and your monthly income will quickly reach 100,000+; various conferences will invite you to participate, broadening your horizons and feeding back into your product updates; with more social opportunities in the same industry, your circle will expand, and you will meet some big shots; you might even find people who recognize you, transform from an OPC, get investment, and become the boss of an AI company.
It is a lever. In an era where the gap between rich and poor and class solidification is increasing, it can allow a small-town student to talk to A10 big shots, and allow a confused young person who just graduated with a monthly salary of 3k to catch the fast train of the era.
Several Tips for Starting an AI Content Creator Account in China (Framework Version)
I've shared a lot of my views on AI self-media above. I believe you now have a deeper understanding. Next, I will share some tips for starting accounts on domestic platforms like Douyin, Xiaohongshu, Bilibili, and WeChat Video Accounts.
1. Account Setup
Do AI self-media in an IP way—the setup of an AI self-media account is very similar to the setup of a commercial personal IP account.
First, we need to come up with a memorable and short name, preferably in Chinese. It can even be your full name or a stage name. (It needs to have a sense of the internet).
Then we choose the priority for the profile picture: Personal photo > Memorable photo > Ordinary profile picture.
Write the personal bio in separate lines. The first line can be your slogan, such as "Sharing AI dry goods, helping more people use AI well~". The second line can be a self-introduction, such as your educational background + work experience "Zhejiang University grad, former ByteDance PM". The following lines can be written freely, such as results achieved or visions.
Due to censorship, try not to leave text on the homepage background image. Choose landscapes, personal photos, or AI images as the background. You can express your attitude or even hobbies through the background image.
Finally, set up the account settings, complete real-name authentication, and turn on/off privacy settings as needed. If you are afraid of family or classmates recognizing you, then block them. I don't agree with some people who say: "If you are even afraid to show it to friends and family, you won't do well in self-media." On the contrary, I believe that the mindset for self-media is very important. If you are afraid of losing face if you don't succeed, then block them, block all sources of anxiety, and then work hard in silence. If you don't succeed, just act like it never happened; anyway, no one knows. Once you succeed, they will naturally know and look at you with new eyes.
Wang Dalang's domestic social media homepage
A good account homepage is just right—not pretentious, not messy, and a bit interesting. Friends who don't know how their homepage is set up can post in the comments, and I will give you a score + modification suggestions.
2. Content & Topic Selection
First, let's talk about content.
I believe the quality of content is determined by information density, speed of dissemination, and platform algorithms. Video > Graphics, appearing on camera > not appearing on camera. Summary: Make videos, and you must appear on camera! (Of course, big shots like Digital Life Kazik are exceptions!)
According to my observation, AI bloggers are currently roughly divided into two schools: the Technical School (AIGC design technical school and code technical school) and the General Traffic School.
The Technical School generally produces horizontal videos, more than 2 minutes long, with many editing special effects and high quality. However, the threshold is high and it consumes a lot of time and energy, but it is also the easiest to start an account. It can be said that at this point in time, as long as the topic selection of Technical School bloggers doesn't go off track, the probability of outputting viral content is 100%.
The General Traffic School generally produces vertical talking-head videos, with a fast pace and strong camera presence, somewhat like a friend's perspective. Sometimes sincere, sometimes funny, sometimes complaining about grievances. Just sharing insights about using AI in life can quickly accumulate followers.
The advantage of the Technical School is that it doesn't require great camera presence or expression skills, focusing on content and editing. The disadvantage is that the IP character is not personal or three-dimensional enough, and the technical difficulty and learning cost of video are high. Follower stickiness is poor, but the relative quality of followers is higher and more vertical.
The advantage of the General Traffic School is that the content is simple and easy to make, followers grow fast, and stickiness is high. The disadvantage is that it requires great camera presence and expression skills, and even a face that fits the internet—not necessarily very handsome, but it must make people feel comfortable and approachable. At the same time, because the content is more entertainment-oriented, the followers may be more general, of lower quality, and not vertical.
Here are some AI self-media bloggers I think are worth learning from: Qiu Zhi 2049, Digital Nomad, Zhang Zala, Muzi Doesn't Write Code, Xuan Jiang, Huang Bai, Absolute Little Dream, Simon Smart Egg, Gui Zang, Dog Brother's Notes. They include both Technical and General Traffic schools across different platforms. You can search and mark them yourselves.
Now let's talk about topic selection
Today I will share two methods I use to catch viral topics.
1. Information Gap Relocation Method
If you are a veteran user of X, you will notice a phenomenon—Why do hot topics discussed on X a few days ago only become popular on domestic platforms today?
If you observe carefully, you will find that a lot of AI information starts from the English circle, then is translated into Chinese, then starts spreading in the Chinese Twitter circle, and then becomes popular in the entire Chinese circle. This time gap ranges from a few hours to one or two months. What we need to do is take advantage of this time to thoroughly study this hot topic and practical methods, and then create viral content for domestic platforms.
This is how I usually do it: First, search for tutorials on hot content on Twitter. For example, if you searched for keywords like openclaw in February-March, or codex in May-June, there would be a lot of high-quality articles. You bookmark them all, feed them to your AI to summarize, and then go to domestic platforms to search for these summarized keywords. If there are already viral posts but they just came out, or signs of going viral are slowly appearing, you can add them to your topic library as backups. Then you create secondary content based on the articles and turn them into your video content.
This method of turning high-quality Twitter articles into domestic talking-head explanation content is what I call the Information Gap Relocation Method.
Don't be stingy with your bookmarks; add high-quality articles to them.
2. Growth Record & Life Integration Method
This method is very grounded and very effective for starting an account.
Step one: Go and see what content domestic AI self-media big Vs have been posting recently. For example, if Qiu Zhi posted a Codex tutorial and the video went viral, then OK, it's simple! We follow this hot topic and record, learn, and use it from the perspective of a beginner in daily life, while showing the results.
The structure of this viral script is roughly: "Family, this Codex is so amazing! Didn't I go for a cross-border e-commerce operation interview the other day? I actually didn't know many knowledge points and didn't know how to write my resume. As a result, I sent the information to Codex, and it helped me get everything ready for the interview in just 3 minutes. It's so amazing! Let me teach you how I did it. Step one xxxxx, then xxxxx, and finally xxxxx."**
Repeat this a few times, and you will have mastered this growth record and life integration method, and it will definitely produce viral content. Because this kind of content gives readers a sense of equality with you. They won't feel that the content you are talking about is difficult; instead, they will think: "Hey, this person's life and income seem even worse than mine, and he can learn Codex. I definitely can too." So they will follow you, hoping to grow and progress together with you. This was the method I used when I first started my self-media account. I used the perspective of a college student graduating and doing cross-border e-commerce, recording and explaining every step from opening a store to selecting products to listing... but actually, at that time, my monthly turnover had already reached $20,000.
Summary of content and topic selection; you can save it and follow it for implementation.
3. Filming & Editing
My filming tips:
If filming vertically, use an iPhone's front camera plus the Kaipai APP. Kaipai provides a teleprompter function, so you can just look at the screen and read. Note that when reading, you should also have camera presence; don't just read like a textbook.
If filming horizontally, you can use the DJI Pocket series. If you have a limited budget, use the Pocket 3. It can provide depth-of-field blur, making the portrait effect better and your video look more professional and high-end. For horizontal screens, you can choose 16:9 (common) or 4:3 aspect ratio (wider, convenient for vertical screen users to watch).
At the same time, the setting should preferably have some depth of field. You can buy a floor lamp for lighting later, and pay attention to sufficient front lighting. You can put some eye-catching or tech-style gadgets behind you. Students who don't know how to set up lighting can learn directly on Douyin, Xiaohongshu, or Bilibili; there are plenty of tutorials.
My editing tips:
Using CapCut (Jianying) is enough. You can use your phone when you're out and your computer when you're home. After getting a membership, everything is synced in the cloud, which is very convenient. Usually, you can accumulate editing materials, such as commonly used fonts, transitions, sound effects, and BGM. Collect them all so you can continue to use them in the future.
If you need computer screen recording and want to achieve that zoom-in/out effect and cursor movement, you can use OBS (high learning cost), FocuSee (paid), or Openscreen (open source but unstable).
Since text cannot demonstrate filming and editing very well, I will write articles to share in detail later, or make them into Skills. You can stay tuned.
4. Censorship Bypass Techniques
I believe many friends on X who go to domestic self-media platforms have a headache with one problem—censorship.
I see many people complaining: What on earth is domestic platform censorship blocking???
I clearly used those word-checking software, and I clearly paid attention to many things. Why can't my content pass the review???
These were also problems that I used to have a headache with. I used a lot of software and paid for many memberships, like Chanmama and Linke word check, but in the end, what didn't pass still didn't pass. Finally, I found that I was focusing on the wrong thing. I was always obsessing over the sensitivity of the content, such as changing 'money' to 'rice' or 'buns', changing 'favorite' to 'fav', or changing 'most' to 'm0st'.
In fact, what is being blocked are the domestic and foreign restrictions on AI content, and even platform restrictions. For example, last year on Douyin, if you posted Alibaba's Qianwen or Tencent's Yuanbao, it wouldn't pass the review, but if you posted Doubao, it was fine.
The most ridiculous case I encountered was during the Openclaw wave. It started getting hot abroad in February, and it exploded domestically in early March. I followed the trend and updated two pieces of Openclaw content in early March, and both went viral. But in mid-March, I shared a full process for installing Openclaw, and as a result, after getting 3,000 likes, the video stopped getting traffic.
Finally, I found that other bloggers, even in their latest content, couldn't find new Openclaw content anymore. What was left was old content, and the traffic stopped growing. Later, I looked at my phone and understood; CCTV News posted on WeChat and Weibo, warning about the risks of Openclaw. But the grand occasion a few days ago was still—many places released 'Ten Measures for Lobster', vigorously supporting Lobster OPC. I went to ask various platform reviewers, and they couldn't give a reason, but I am certain that was the reason.
So, sometimes passing the review depends on luck. You must keep a calm mind~
Currently, the AI content that I've summarized as being able to pass or not pass is as follows:
Can pass: Codex, Claude, Hermes, ByteDance products and models (Doubao, Trae, Jimeng, Coze, Seedance, Seedream, etc.), Tencent products (Yuanbao, Marvis, Workbuddy, etc.), Alibaba products (Qianwen, Wukong, Qoder, Mulerun, Ant Afu, etc.), Minimax products, Kuaishou models (Kling), various domestic Claws, Kimi products, Zhipu products, StepFun products, Baidu products, and products from various small domestic AI factories, etc.
Cannot pass: Image-2, Opus, Openclaw, GPT, and other foreign AI models or Agent project names are very easy to get stuck in review.
Another small tip: try not to make comparisons in your content. For example, if you say "Codex is much better than Trae!", 99% of the time it won't pass the review.
5. Commercial Monetization
Currently, the monetization of AI self-media accounts is very diversified. I have roughly summarized several methods and given some suggestions for each.
1. Accepting Advertisements
This is currently the fastest and most considerable way to monetize AI self-media. Brands have different placements based on the user profiles of different platforms. Current AI advertising budgets across platforms: Douyin = Xiaohongshu > Bilibili.
Brands are the 'gold masters'. Sometimes they allocate budgets directly to their own BD departments for PR. But some will also directly package it to media companies to connect with bloggers. At this time, some concepts will appear, such as 'above water' and 'under water': which means going through the platform's official advertising channel or not. For example, Douyin's Xingtu, Xiaohongshu's Pungying, and Bilibili's Huahuo. Going 'above water' requires a platform commission and advertising review, but the advantage is that it is more compliant, and brands can run information flow ads. 'Under water' means bypassing the platform and publishing directly, but sometimes it will be judged as an advertisement, thus getting stuck in review.
Another concept is 'rebate': which is the benefit fee taken by those media MCN companies or independent PRs. For example, if your quote for one video is 20,000 RMB, and an AI application finds you through these intermediaries, and you have no way to directly connect with the brand, you have to give them a certain rebate. If the rebate is 20%, then the rebate is 20% * 20,000 = 4,000 RMB in benefit fees before they are willing to cooperate with you to submit the project and push it forward.
Currently, there are many advertising brands in the AI industry with sufficient budgets, ranging from software to hardware. Yesterday at WAIC, I also communicated with the head of the marketing department of a mid-tier robot company. Their monthly placement budget starts at 5 million RMB. Currently, the AI industry advertising is still a seller's market, meaning bloggers have the advantage, and we can choose which ones to accept. After the brand or PR sends you the Brief document, you can try the product first. If it feels good, accept it; if not, you can politely decline.
As long as you are not starving right now, remember this: If the content is good, there is no shortage of advertising brands.
2. Knowledge Payment
I have always believed that monetizing IP through knowledge payment is a double-edged sword. After visiting Lu Damuo last month and chatting with him, I became even more convinced of my idea.
Because knowledge payment generally explodes at the end of an industry (except for success studies, emotions, and numerology-type schemes), which means the people who should make money in this industry have finished making it, and the remaining people who master the methodology can't make much money, so they can only earn the last scraps of the industry by selling shovels.
And your delivery might be very heavy, but there are many fools, which leads to a problem—even if your knowledge payment content is very good and high-value, there will always be people who don't understand or have poor execution. After trying it and not getting results, they will think your product has no value, but you think you've done everything you should, leading to a dead end. In fact, knowledge payment is not like selling physical goods; returns don't matter because they can be sold again. But knowledge payment is selling time and energy, and even if you sell recorded content, you can't avoid being freeloaded.
Therefore, I found that either do knowledge payment smartly, or do it only when your IP is large enough, your delivery is light, and 99% of customers will be satisfied with the product (such as AirCity's editing course, which sells for dozens of RMB but is of very high quality). I'll expand on this in a moment.
The AI industry is a slow-bull industry. Its narrative is grand enough, and its knowledge payment life cycle will be very long, from Tsinghua Li Yizhou's 9.9 RMB courses to various Doubao and Jimeng courses for the elderly, to various AIGC and short drama teaching courses, to high-quality Vibecoding knowledge payment content, and even systematic robot-related courses in the future. The coverage and content of knowledge payment in the AI industry are incomparable to other industries.
Okay, it's clear that the AI industry can indeed do payment. So how to do payment smartly?
One secret is—selling courses under the guise of not letting them sign up for a course.
- Offline Gatherings
You can call everyone to come in the form of a ticket, and then distribute your free courses. In fact, your course fee is included in the ticket, and the advantage of offline is that it can bring customers closer to you and better provide emotional value. "The emotional value is full, you've added me on WeChat, and we've become friends. Brother, it would be too much for you to refund the course now, right?"
At this point, don't you think you can do a second conversion offline? Start selling your 19,800 RMB course?
Wrong!!! What you need to do is preheat offline first, telling everyone: "Hey brothers, recently I want to do a very high-value AIGC teaching, or recently I did a good project selling Tokens. At that time, I might just find a place for a few days, and we can run through 0-1 together. What do you think of this idea?" Then let those who are interested add you or DM you, and then convert them individually online. Get it?
- Member Groups & Douyin Membership
You can actually achieve knowledge payment through a price threshold, whether through live streaming or converting public traffic to private traffic, and then continue to sell high-ticket products after they reach private traffic.
Many financial and business bloggers do this. For example, if you join my Douyin membership, you can enter the member group, where I share good sector directions and good projects every day, plus some free materials, etc. In fact, this is already the first conversion. You entered the group and paid a membership fee of 199 RMB per month, which is essentially no different from signing up for a 199 RMB course.
But you might say: "This is different! I spent 199 to enter the big shot's circle and communicate with everyone. I'm not buying a course!"
Haha.
The advantage of this method is that it can also help you expand your community. In the future, if brands want to advertise with you, and your private traffic reaches 50,000 people, the quote won't be low! And if you create any products, the people in the community will be your first batch of users.
It's like winning the lady and hoarding the soldiers at the same time.
3. Product Monetization & Resource Monetization
Actually, I wrote about this earlier. Why do we need to do AI self-media? Because we need to seize the users and traffic of the AI era. Once we have traffic and users, we can use products or resources for monetization. We can develop products ourselves or form a small team to develop products, using followers and the community as our users and a place to collect feedback for continuous iteration and optimization.
When the product is almost ready and the data is up, we need to do the social media for the product well, both domestically and abroad. At this time, we can prepare a BP, talk to VCs, and socialize in various occasions. Then, with money, everything goes smoothly~
This monetization method has the highest ceiling but is also the most difficult, because many uncertainties may arise in the middle, and it's hard to see returns in the short term.
But fortunately, we have income from advertisements or knowledge payment to support the product. This is also why we need to do AI self-media. If you think about it carefully, AI self-media is like a six-sided warrior! You can see money in the short term and achieve freedom in the long term.
Summary of commercial monetization; you can save it and follow it for implementation.
Conclusion
Writing up to here, I feel this article is a bit too long and the information density is very high. But for those who can read to this point, I feel that your success in AI self-media is almost certain!
Because you have already mastered the entire framework of doing AI self-media. Later, you just need to have full execution and follow this Roadmap I wrote.
Of course, the journey of doing AI self-media is certainly not smooth sailing. There are many small details and sticking points, such as posting N videos without traffic, lacking camera presence, taking 3 days to edit one video, facing cyberbullying and insults, or no brands looking for you, making it impossible to earn money... all kinds of difficulties can easily make you give up.
But I hope that one day when you want to give up, come back and read this article again. Please remember that there was once a person named Wang Dalang who started seriously just like you, and once accompanied you to fight side by side on the road of AI self-media across the screen!
At this moment, we might still have to grit our teeth and persist in days when no one cares.
But I believe that the light boat will eventually pass through the ten thousand mountains.
When you and I have both become million-follower AI bloggers, at that time, we will meet at WAIC and greet each other with a smile:
Finally meeting at the top! Fortunately, I didn't give up back then, hahahah (the two look at each other and laugh loudly).
Finally, I wish everyone a smooth takeoff on the road of AI self-media~
In the next two days, I will also establish an AI self-media exchange group to share dry goods in the group.
By the way, what would be a good name for the group...?
Walking through WAIC... playing around a bit~
