Author: One.
Editor: Meimei.
People seem to prefer AI as a toy rather than a tool.
Recently, an AI-native app called Coaxing Simulator went viral, attracting 600,000 users in one day.
The coaxing simulator sets up a variety of couple quarrel scenes, whether you are a man or a woman, you can let the AI play your boyfriend or girlfriend here, and according to your low EQ or high EQ reply, the system's "forgiveness value" score will rise or fall, in order to evaluate and exercise your coaxing ability, it can be said that let the AI teach you how to coax the other half.
Relying on the gamification experience, it attracted 600,000 users in one day.
This explosive coaxing simulator product is actually inspired by developer Wang Dengke's own real life experience.
He posted that when he quarreled with his girlfriend a few months ago, when he saw the other party's angry appearance, he would imagine her as a robot, and a virtual progress bar would be displayed on his head, and his response to the other party would affect the change of the progress bar, so this became the product logic of the coaxing simulator**.
The first to catch fire was actually the web version of the coaxing simulator.
As you can see, it offers many different dialogue scenarios, and unlike some pure chat or role-playing AI dialogue products, the Heaxing Simulator adds some gamified experiences, such as leveling up and scoring systems. In different dialogue scenarios, the forgiveness score will be displayed at the top, and the higher the score, the more likely it is to coax the other person to avoid quarrels.
In addition to some preset scenes, the coaxing simulator also supports users to set their own communication scenes: after filling in the scene information and dialogue samples, they can see the scenes they have created. The gamified experience aroused a lot of interest, and developer Wang Dengke actually tried it as a hobby at first, but due to the posts in QQ space and the spread of QQ groups, overnight, the number of users of the coaxing simulator exceeded 10,000, and the daily active users were close to 100,000, and then more than 600,000 users used the product.
The popularity of coaxing simulators seems to be due to accident, but we can actually get a glimpse of the internal logic of some popular products in the AI era.
Gamified AI is easier to get out of the loop than practical AI.
In fact, the coaxing simulator is not the first AI product in China. Before it, there have been a number of AI products that have become popular.
In July last year, a small program called "Wonderful Duck Camera" went from obscurity to suddenly hot.
You only need to upload 20 faces on the applet**,You can generate your own "digital clone",Then choose a template,You can form a variety of photos,Ancient、Professional、Modern、Korean、Japanese,You can generate,And through refinement,Modify you to any desired look,And this set of operations,Only 9 pieces 9,Compared with professional photo studios,It's equivalent to no money。
All of a sudden, many people's circle of friends were swiped by the photos generated by the Miao Ya camera. So much so that the Miao Ya camera is called "the first subversive hit in the AIGC era".
And after the Miao Duck camera, in October, with amazing ** such as "Guo Degang speaks English crosstalk" and "mildew speaks Chinese", an AI named Heygen became popular.
During that time, many celebrities and Internet celebrities have become "foreign language experts", speaking English is no longer nothing, and Japanese, Korean, and Spanish are proficient, as if the translation industry is about to be subverted, and the language barrier between different nationalities is about to be eliminated.
In the following November, there were many AI-related challenges on the hot list of Douyin, a domestic short ** platform. For example, related topics such as "AI American-style ID photos" and "AI enlarged pictures that make you unexpected" continue to dominate the list, and there are also a lot of hilarious AI**. After that, the use of DALL·E3 in ChatGPT to generate some outrageous ** has also formed a popular trend of "AI chase and kill order" on Douyin, and there are many** more than 10,000 likes.
And in December, a "Terracotta Warriors and Horses Dancing Subject 3" ** attracted a lot of attention on Weibo, and it was realized by using the function of "National Dance King" in Alibaba's Tongyi Qianwen app. Upload a full-body picture of the character, and then select the dance, and the AI can make the character in the dance dance.
In the six months after the explosion of the Miao Duck camera, it can be said that almost every month, there is an AI-related gameplay that has gained a lot of popularity on social **, in addition to the above, there are also AI patients of Zuoyi GPT and so on. The popularity of these AI products seems to be accidental, but in fact, they are all AI with entertainment attributes and gameplay, which can attract a large number of users in a short period of time.
You don't have to go head-to-head with big manufacturers, and AI entrepreneurship may have to change your thinking.
Since ChatGPT set off a wave of AI, many startups have emerged in China. As of the end of last year, in just 8 months, 238 large models were born in China, and the expression of "100 model war" is no longer accurate, and it should be called "200 model war".
However, among the 238 large models, the problem of homogenization is serious, and chatbots are one of the "hardest hit areas". Wenxin Yiyan, Xunfei Xinghuo, Baichuan Model, Zhipu Qingyan, Doubao, etc., are obviously very "volume".
Since the beginning of the AI wave, large models have always been regarded as a tool for enterprises to reduce costs and increase efficiency, and the introduction of large models into the workflow of enterprises can achieve personnel streamlining and efficiency improvement. Therefore, if you are competing on the AI track with strong tool attributes, then the dimension of competition between different large models is who has stronger and more practical functions, and it will inevitably fall into the competition of computing power, algorithms and underlying resources of data.
Because the training of large models needs to rely on professionals to write algorithms, and it needs to spend a lot of money to purchase data and computing resources, the investment needs to be very large before it succeeds, and it is actually difficult to get enough investment on the track with serious homogenization. So is the entrepreneurship of ordinary people in the field of AI doomed to failure? Not really.
The popularity of AI with entertainment attributes such as coaxing large models may mean that at this stage, C-end users actually prefer to use AI to "increase laughter" rather than letting AI reduce costs and increase efficiency. The above-mentioned Heygen, AI Extended Picture and AI Patient, etc., their initial functions are actually practical, but they are out of the circle because of their entertainment attributes. This may give AI entrepreneurs some new ideas, that is, to develop AI with entertainment attributes, rather than AI with tool attributes.
Entrepreneurs do not necessarily have to implement the concept of "using AI to reduce costs and increase efficiency" to the end, but can achieve some richer and more unique entertainment gameplay through AI. The reason for this shift is that AI is currently more like a toy than a tool for most ordinary users. Moreover, compared with the tool attributes of reducing costs and increasing efficiency, the development direction of entertainment attributes is more, and the corresponding competitors will also be reduced.
Moreover, entertainment attribute AI does not pursue accurate and perfect results. On the contrary, the appearance of unexpected results can enhance its entertainment attributes. For example, the "AI Expanded Picture" in the Douyin gameplay, in which the high likes are the unsatisfactory expansion effect, and even the outrageous AI**, as well as the AI patients in the GPT of Zuo Yi, it is also because of some unexpected answers in the dialogue process, which can stand out from many role-playing dialogue AI. The wonderful duck camera and coaxing large models, etc., are limited templates or scenes, and the content that needs to be processed is less than the infinite content generation, so the consumption of underlying resources such as computing power and algorithms can be reduced for entertainment AI, which means that ordinary people can invest less money in entrepreneurship.
How to monetize with entertainment AI?
When we change our thinking, the next problem that needs to be solved is how to monetize this kind of entertainment AI.
The computing power resources of large models are tight, so a unique charging model was born.
For example, heygen is free to use, but there is a limit to the number of times. And due to the problem of computing resources, the generation needs to be queued, and paid members can "cut the queue" to get the generated content faster. This model is actually worth learning from many entertainment AIs, because it takes a lot of time to generate a lot of AI content, which can completely guide users to reduce time costs by paying.
For example, the wonderful duck camera is hot at the time, even if you have paid 99 yuan, you also need to wait for 5-7 hours, and Tongyi Qianwen's "National Dance King", it also takes dozens of minutes to generate**, in this case, users can be allowed to pay and get the desired effect faster.
Similarly, we can also learn from the charging model of the gaming industry. At present, the coaxing simulator has more than 10 free replies in each scene, and there are two payment models, which are charged weekly and monthly, with a weekly fee of 6 yuan and a monthly fee of 12 yuan, which is similar to the monthly card model in the game industry. And the wonderful duck camera is the first generation itself requires a pre-charge 99 yuan, to get a high-definition picture, you need to use diamonds, which can be obtained by sharing or purchasing.
In addition, it is also possible to unlock a wider variety of models, characters, and scenes by paying for them. For example, more photo templates have been introduced in the Miao Duck Camera, more dance movement models have been introduced in the "National Dance King", and more characters other than AI patients have been unlocked in the Left Doctor GPT, etc.
Heygen's explosion allowed the company to get $5.6 million in venture capital in November last year, which is enough to give entrepreneurs some confidence. There are many monetization channels for entertainment AI, and AI entrepreneurs need to design a payment model based on their own AI products or gameplay, so as to bring users an entertainment experience while realizing monetization.
Today, with the increasing maturity of AI technology, small businesses founded by ordinary people do not have to compete with large manufacturers on the traditional track. By developing AI applications with entertainment attributes, small businesses can find their niche in the market and differentiate themselves from the competition.
These apps are not only fun for users, but also for small businesses to grow in. Through innovative business models and monetization strategies, small businesses can carve out a unique development path in the AI field to achieve sustainable development.
With the continuous advancement of AI technology, we have reason to believe that more creative and interesting AI applications will appear in our lives in the future, bringing more surprises to users.