While you're still sharing AI-generated images online, some people are already making money from writing prompts.
A good line of talk is worth
ten filters
Some time ago, LeBron James visited China. I saw on my WeChat Moments that several of my friends had taken photos with him. I was wondering why they didn't go to work and went to the event. Only after asking did I know that they used DreamAI 4.0 to generate these photos to ride on the wave of popularity. In the world of office workers, there aren't so many beautiful spur - of - the - moment adventures.
Well, needless to say, I immediately went to the official website of DreamAI to create one myself:
● A group photo with a celebrity generated by DreamAI 4.0
Those who are interested can also give it a try with this prompt ——
Fuse Image 1 and Image 2 into a single overhead selfie of two people. The composition of the picture should be compact, with the two main subjects standing close to each other. Their heads should be tilted slightly upwards, and their eyes should look directly at the camera to create a strong visual impact. The person on the left should stand slightly forward. Keep the same pose and characteristics as the main subject in Image 1, and maintain a high similarity in facial features. The person on the right should keep the same pose and characteristics as the main subject in Image 2, and maintain a high similarity in facial features. The body of the right - hand person should be slightly tucked in. The shooting angle should be a high - angle overhead shot, exaggerating and enlarging the head proportion, in line with the typical Japanese - Korean visual selfie style. The background should be pure white, simple and clean, to further highlight the human subjects. The picture style should lean towards the Japanese visual style, with high overall picture clarity, as if taken with the front - facing camera of an iPhone. The final result should present a delicate, fashionable, and slightly edgy group photo effect. The people should blend seamlessly into the picture, with a natural visual transition, and the overall picture should be brightly and evenly lit.
I didn't come up with this set of prompts myself. I found it online. Nowadays, these photo - editing prompts are everywhere on the Internet. You can find a large number of them with a simple search, and any random post can get hundreds of likes. Some people even make a business out of it. Many stores sell prompt templates for just a few yuan, and it's also a considerable income.
● Posts related to the topic searched on Xiaohongshu
Doesn't this scene seem familiar? Just like the suddenly popular Nano Banana at the beginning of the month, overnight, everyone was showing off group photos with celebrities, cartoon - style avatars, and 3D figurine photos. Their common feature is: they look incredibly real at first glance.
Actually, DreamAI was also popular for a while before. People used it to edit photos, change various backgrounds, lighting, etc. But the impact brought by DreamAI 4.0 is even stronger. Why?
Let me share my personal experience first. The pictures generated by the previous DreamAI 3.x versions, although they could make people laugh, still had a bit of an “AI feel”. It's hard to put into words, but the faces didn't look real and couldn't blend into the picture. Especially the eyes, they either looked dull or unfocused, and it was obvious that they were “fake” people.
But with DreamAI 4.0, this fatal shortcoming in the eyes has almost been overcome. There is light in the pupils, and the expressions carry a bit of casual emotion, even creating an illusion of being accidentally captured.
● The left one is DreamAI 3.1, and the right one is DreamAI 4.0
When you post the photo on your WeChat Moments, people's first reaction won't be “This photo is well - edited”, but “Is this real or fake?” This blurred line between truth and falsehood is the perfect breeding ground for spread. The reason why DreamAI 4.0 can make such a big breakthrough is that it has achieved the ultimate in “speed, accuracy, and authenticity”.
Speed: It has a unified multimodal architecture, which enables one - step understanding, generation, and editing without switching between different models.
Accuracy: With a consistency algorithm, it locks in the core features of the same person and maintains stability from different angles.
Authenticity: It can generate 4K high - definition images with details so real that you can see the texture of clothing and the skin's texture.
Some people may ask: This function was already available in the previously popular Nano Banana. Why has it become so popular with DreamAI 4.0? It's so popular that even my dad has started using it to dress up my niece and take group photos with various celebrities.
In my opinion, it's the result of several factors combined.
Firstly, the technological breakthrough has reached a critical point. The previously generated pictures still had a bit of a plastic feel, but this version upgrade has just enabled “virtual group photos” to cross the threshold of “looking like a real person”. Just like a common saying in the photography circle: Once the eyes connect, you believe it.
Secondly, DreamAI 4.0 can be used within Doubao. Compared with Nano Banana, this greatly lowers the usage threshold. Moreover, Doubao has a monthly active user base of 157 million, currently ranking first among China's native AI apps.
Thirdly, celebrities naturally add a lot of traffic. If you post a photo with a ancient - style model, people may just give it a like. But if you take a group photo with a top - tier celebrity, it's a whole different level. Celebrities are social currency, and AI just makes this currency accessible to everyone.
What's even more interesting is the prompt trading behind AI - generated images.
In recent days, on social media and second - hand platforms, I've seen people selling “collections of prompts for group photos with celebrities”. The prices range from a few yuan to dozens of yuan, and some even offer “customized prompts”. You just tell them who you want to take a photo with, the scene, and the clothes, and they'll write the instructions for you, guaranteeing a well - looking “scandal photo”. The prompt I shared above was found on a social platform.
When I saw this, it suddenly seemed very familiar to me. Isn't this just a new version of the PS filters and Lightroom presets back then? In other words, it's called the “Prompt economy” now.
● Search for “DreamAI prompts” on Xiaohongshu
In the era of AI creation, prompts are the new filters. A good prompt can help you directly generate magazine - quality images. A poor prompt, even with the same model, will result in an obviously fake - looking image. Ordinary people don't have the energy to spend hours adjusting parameters, so they'd rather spend a little money to buy a ready - made “universal key”.
This in itself has become a small business.
But I think what's truly valuable isn't just a few fake group photos on WeChat Moments, but the trend behind it: We're entering an era where “what you think is what you get”.
Previously, we needed a photo studio, lighting equipment, makeup artists, and post - production editors to take a group photo comparable to those in celebrity magazines. Now, just a single prompt is enough.
You can take a photo with your 18 - year - old self;
You can also write a prompt for your future self and see if you'll still smile calmly at the age of 60;
You can even generate a life segment of “what if I had gone to study in Paris”.
DreamAI 4.0 isn't just a photo - generating tool. It's more like a “life generator”.
It allows us to vividly see what we might have been like on the paths we didn't choose.
So, when I see a friend's sweet group photo with a celebrity on my WeChat Moments again, I won't instinctively be nosy. Instead, I'll just think: “Well, this prompt is well - adjusted.”
As for DreamAI 4.0, it's just a stage - specific product. In the future, we'll see more moments when “what you see may not be real”. In this process, some people will be infatuated with the false appearance, while others will create true meaning. The difference is that the tools are already in our hands, and it's up to us how to use them.
Written by | Ma Yunan
Edited by | Ba Rui
Cover image source | AI - generated
