The AI technology that has been hotly discussed and always far away has finally “successfully landed” this year, and the turbulent ChatGPT has become the gathering place for the enthusiastic attention of Internet giants.
Emerging topics such as ChatGPT, AI painting, AIGC creation, and large-scale models emerged in the AI storm, which attracted everyone’s attention. In March 2023, there will be no less than 10 hot search topics related to AI, which shows that everyone’s attention to AI remains high.
At the same time, there has also been a wave of AI entrepreneurship in the industry. Whether it is a major company such as Ali, Tencent, and Byte, or technology giants such as Kai-Fu Lee, Wang Huiwen, and Wang Xiaochuan, they have all announced the end of AI entrepreneurship. And these entrepreneurial fevers have driven AI-related positions to become hot cakes.
In the arena of China’s AI industry, there are such a group of players who are born in the grassroots and are little known, but spread AI products and solutions in every corner of the streets and alleys.
When the AI ”unicorns” lost money, they had already made a lot of money; when the technologists “held a hammer to find a nail”, their plans had spread among the people, bringing the initial enlightenment of artificial intelligence. However, when problems such as face abuse and data privacy infringement frequently arise, they have also become the group that rushes to the forefront and is closest to the “dividend”.
A businessman with a low level of realization of ideas: Lao Ba – “After sharing with 3,000 people, someone directly asked me how to ‘realize'”
No one would have imagined that the first batch of people who made a lot of money from ChatGPT were neither R&D personnel nor major domestic manufacturers, but sellers of tutorials.
A few days ago, a form of “teaching people to make money with ChatGPT to make millions a week” went viral on social platforms. The revenue estimates of the top 5 are all over one million yuan, and the data comes from a knowledge payment app.
However, the people at the top of the table are relatively well-known knowledge bloggers and have a certain fan base. For example, at the top of the table, “ChatGPT AI Monetization Circle”, which relies on selling the writing ability of the official account, has been in operation for 42 days, with 11,900 members and a price tag of 99 to 365 yuan.
Similar to this, there are endless courses that teach you how to “get rich with AI”. The theory that “it is better to sell shovels than dig gold mines” has been mastered by people.
Lao Ba can be regarded as an old user of major knowledge payment platforms. His three most direct monetization methods rely on the most advanced AI technology: selling courses, selling solutions, and collecting membership fees for community building.
Beginning in October last year, Lao Ba created a Xiaohongshu account for the English education track, and the number of fans rose to 140,000 within 100 days. Therefore, I also made a sharing in a community of 3,000 people called “ChatGPT, the creative tool of Xiaohongshu”.
Where there is a market and demand, there will be a provider. A common need is that everyone wants to make money. People kept coming to ask. At that time, Lao Ba only shared his experience after using it. After posting it for a day, more than 100 people came to add friends and ask for advice.
The way Laoba earns money indirectly is to post his experience and tutorials in the community of the paid platform, and there will be a reward mechanism in the community to share his experience for a fee, so he earns five easily. number.
Later, Lao Ba’s courses became more and more abundant, and the knowledge he shared became more and more dry. “It is always said on the Internet that AI will eliminate 90% of people in the future, but in fact, in the future, the 10% of people who have mastered AI will be eliminated. 90% of people who have not mastered AI” “You must pay, which represents your determination to embrace and explore the future” “ChatGPT+ official account explosive writing, starting from 0, reaching 100,000+ in 7 days”.
The beginning of the course is to introduce the most basic things such as what is GPT and how to register a GPT account. Comparing the two courses, it is obvious that 365 yuan will be more detailed than 99 yuan. For a content polish, 3 possible usage methods are listed. In addition to basic registration tutorials, the two courses also have advanced tutorials like how to improve efficiency, tune GPT, and even predict lottery tickets.
However, in comparison, in terms of content richness, 365 yuan is indeed more abundant than the 99 yuan tutorial. The 365 yuan tutorial also shared a lot of plug-ins, AI small websites, etc.
What are the characteristics of the To B industry? Personalized customization; long customer acquisition cycle (longer decision-making process); product implementation costs; linear growth; value sensitive. This is Lao Ba’s most direct feeling.
There is a saying in the AI monetization circle where Lao Ba is located: “everyday the industrial revolution, every night the Renaissance”. There are endless products in this circle. Ideas, ideas, and all kinds of information are changing with each passing day, so use the best and top tools. Personally, I don’t have time to understand all the big models. Tongyiqianwen and Wenxinyiyan have all tried it, but they have not developed a habit like using ChatGPT, throwing in everything and asking about it.
Lao Ba used to be a product manager, but after earning money, he resigned and worked for himself. According to his judgment, the development of technology will lead to the loss of human self-worth, and the “hollow heart disease” will become more and more serious. So doing self-media now, I also want to help people find their own “works” and regain their real, non-instrumental sense of self-accomplishment and sense of value.
AI may be the beginning of a “liberation of tools for thought”.
“Visible business” light user: Leek – “After fixing photos for my girlfriend, I resigned and took to the streets”
At first, Leek just thought it was fun and made people happy, so he did it.
Leek has learned to draw since he was a child, and his main job is advertising design. When the AI era came and smart photo editing software came out one after another, Leek was very excited, thinking within a day, “Can I use it to restore old photos from before”? In less than a week, he figured out how to play the software.
When he posted his first restoration work on social platforms, the data exploded. Some asked him to start a class, some asked Leek to help repair photos, and some companies asked him to join the company’s IP plan and asked him to specialize in projects.
The primary market expresses optimism by spending money like water. AI companies make no secret of the nobility of technology labels in publicity. It seems that the era of a set of SDK sweeping the world is just around the corner.
Leek knew that this was some potential business opportunity, but he didn’t reply. It would be embarrassing if the other party offered an unrefusable price.
Later, Leek felt that this would be a market with huge potential and a lot of demand, so he came up with the idea of starting a business. On the one hand, with the development of digital technology, more and more old photos have been preserved, but due to a long time or improper preservation, the photos have problems such as damage, blur, and discoloration. These problems will affect the value and appreciation of photos, and there is a demand for restoration of old photos in the market, and there are still many people who do not know how to use restoration tools. Therefore, old photo restoration is a potential market.
But at the same time, there are certain technical thresholds. Restoration of old photos requires mastery of image processing, color restoration, restoration materials, etc. Although there are some repair software and tools that can help restore old photos, the repair effect still requires manual fine-tuning by leeks.
In the first month, Leek took a notebook and rode a small car around the streets, and started working as a part-time job. The profit margin of old photo restoration mainly comes from restoration costs and related product sales. The cost of restoring an old photo generally ranges from tens to hundreds of yuan, and the sales of related products such as postcards, albums, T-shirts, etc. can also bring additional income to the restorer. In addition, according to market feedback and user needs, Leek has also developed some additional products to further expand profit margins.
After exploring, Leek extracted the complete process of this business by himself.
Selected crowd: middle-aged and elderly people around the community, towns and other places, as well as the student group near the school, who have needs and feelings.
Select repair items: Select the old photos to be repaired, and determine the specific repair items, such as color restoration, aging and fading repair, scratch repair, etc.
Preprocessing: Preprocessing old photos, such as removing stains, repairing damage, etc. This step may require the use of some image processing software and tools; Restoration: Use old photo restoration software or tools to restore old photos, such as color restoration, repair damage, etc. The process of the repair process may take several attempts and adjustments.
Post-processing: Perform post-processing on the restored photos, such as adding text, patterns, etc., to enhance the appreciation and value of the photos.
Save and Share: Save the repaired photos to your computer or other storage devices, and share them with those who need them. Some restorers may also make related products from the restored photos, such as postcards, albums, T-shirts, etc.
Before the advent of smart phones, taking pictures and photography was the “privilege” of a few people. Now, everyone can enjoy the fun of taking pictures, but there is still a vacancy for photo restorers.
Leek has also joined some AI painting groups. The current monetization methods of “making money with AI” in the group are nothing more than taking photos of e-commerce models and making low-end web pages. However, the biggest drawback of many heavy-duty solution models is that you What can be done, others can also do it, which leads to a great reduction in business profits, and most of them need to be driven by relationships in the end. The heavy-duty solutions industry they entered with a bang and a bang can lead to a profit dead end if they don’t pay attention. Looking at the group operations and comparing his current monthly income of more than 10,000 yuan, Leek is full of confidence in this industry and the sinking market.
“I can understand that many people are eager to realize AI, but finding the cognitive gap of the consumer group is the correct path.” So Leek very much agrees with a sentence: AI is like a thousand horses and horses. It needs a king to command where to fight, otherwise it is just a group of stragglers.
Therefore, Leek often tells latecomers who want to ask for their own business experience, “Don’t be anxious anymore, and be the king of AI.”
I “feed” deep AI users in the county seat: Xiao Kui Kui – “The company has newly set up an AI department, and you can make money with your fingers”
Two photos of road conditions filled the computer screen, and the main character was the road sign in the picture. What Xiaokuikui has to do is to compare the changes in the road signs of the two pictures, and then mark them accordingly. Judging, labeling, and submitting… Xiao Kuikui can complete it in a second. In one day, she has to label about 14,000 groups of pictures.
Beside Xiaokuikui, 32 employees are sitting in front of the computer, new pictures are constantly flashing on the screen, and the sound of typing on the keyboard keeps coming and going. Their common identity is “artificial intelligence trainer” and is responsible for data labeling. Outside the office window, it is a typical rural scene. Before the arrival of spring, the vegetation is sparse and the loess is exposed. Occasionally, donkey carts pulling goods pass by the asphalt road in the county town…
Since December 2019, Xiaokuikui’s company has settled in a small town in the northwest. More than 170 county employees work here. Before that, their identities were mothers and farmers in the county. With the increasingly fierce competition in the AI field, these county AI trainers who are at the forefront of the supply chain have become the focus of the Internet.
On the other hand, through this data labeling company, people in the county are connected with the most cutting-edge technology in the world. In the advertisements of the county TV station, the county town is described as “a low-end version of Silicon Valley”, calling on migrant workers to return to their hometowns for employment, and shouted slogans such as “Junior high school education can do high technology” and “Only you can save it”.
The responsibility of data annotation is that among the large number of pictures that people provide to AI, it is impossible to have only one apple. It may be a picture of a fruit basket, a picture of a fruit tree, a picture of a fruit stand, or a picture of a person holding an apple. , or a picture of a dining table. This requires data annotation to circle the apple in the picture, and then it can be used as training data for AI to practice recognizing apples.
Of course, most of the actual data labeling is much more complicated than labeling the apple in the picture, but it is basically not beyond the level that ordinary people with certain experience or common sense of life can understand and master, or they only need a little training It is possible to engage in such data labeling work.
Talking about the data labeling industry, Xiao Kuikui is a little proud, “As much as there is artificial intelligence, there is as much intelligence. New technologies in China and even the world, such as ChatGPT, may not have appeared in society, and we (annotators) are already working for It’s ready.”
To put it simply, the three low costs are achieved to the extreme: low average labor cost, low operating cost and sales cost, and low marginal cost after production expansion.
For the locals, “artificial intelligence trainer” is a completely unfamiliar vocabulary. An “explosive” news spread in the county: a company employee received a monthly salary of over 10,000 yuan.
Regarding the data labeling, Xiao Kuikui learned from the book that two cabinets were filled with the books he bought in the office. The world in the book has brought shocks to Xiao Kui Kui: the Turing test to test whether the machine has human intelligence, the AlphaGo that defeated the top chess player Ke Jie, the ChatGPT of the storm network, Xiao Kui Kui marveled at the changes in the world.
The business of the county AI company has changed from the basic conversion of listening to voice annotations to texts, to annotations in medical, legal and other professional fields that require training to get started. Up to now, facing unfamiliar English texts, employees can only check each letter one by one. Many people chose to leave because they couldn’t do it.
In the county seat, Xiaokuikui feels that she is a “devoteer” behind artificial intelligence products. He and his colleagues didn’t know what ChatGPT was all about. They only heard that this product was very powerful and popular. At the end of the chat, they concluded with one sentence: “No matter how great it is, it needs people like us to mark it behind.” Through With AI work, more county residents like Xiaokuikui have their own lives and worlds.
“AI has actually liberated part of the productivity. Those companies that have AI to work and just fire people are quite short-sighted,” Xiaokuikui said.
It is undeniable that at present, in any field, the top and bottom must be the first to run out, and the middle is often the most embarrassing. At present, the industry is in a stage of “starting to use the top and bottom layers, and the middle layer is still waiting and watching”.
What should the business model of artificial intelligence look like?
Every time I talk about this question, the answer is divided into two extremes: either homogeneous, or unrestrained and unconstrained. The layman shouted shockingly, and the insider denounced him as having no common sense.
And the artificial intelligence itself answered this way:
In contrast, in just a few years, the low pressure in the AI market has begun to spiral into shape.
In just a few short years, some people couldn’t keep up with the hot spots, couldn’t grasp the core, and followed at the end of the team. Some people carry the heaviest burden, gnaw on the hardest bones, without fear of sinking, and make a fortune silently. Some people hand over their guns, run away, run wildly, and try their best to reach the turning point first. Some people shyly replied, not knowing whether to break the cauldron or stick to it, their hands pulled each other, in a dilemma.
No matter what kind of posture they make their debut, enterprises that fail to find a place to stay will be washed away by the big waves. The business war is cruel, and no one takes pity on the lying rose. Only in the cracks of history, tossing and turning, living to the next era.
But the iron law of the market will destroy the order, and it will also rebuild the rules, giving birth to a new Vientiane.
Now the entire industry is still in a stage where “the starting gun has not fired in the 100-meter race and is still posing”. The wave of AI is triggering a new round of social evolution, and everyone will complete a fair race through business path selection.
This is not only a key node for an AI company to gamble its life, but also an excellent opportunity for an AI company to start its establishment in the new era. Those who are the first to adapt will receive the bonus from the era.
In the mirror image of AI commercialization, who can be the king and who is the loser?
A new story has begun.
Let us pick up the binoculars and start a fantastic drifting of a wave of great change along the curve of the times.