After All in AI, why did Meitu fail to make a wonderful duck?

Source: AI Lanmeihui (ID: lanmeih001)

Author: Yan Ye Editor: Wei Xiao

Miaoya AI camera’s out-of-the-box seems to be overnight.

Only need to spend 9.9 yuan, and provide a face, no need to pose, no need to go outside, you can get a set of high-definition refined photos, and at the same time, you can generate “Datang Princess”, “Suit Elite”, “Suit Elite” according to user preferences. “Ancient Chivalrous Man” and many other styles of portraits.

**This price-performance ratio, no one will say “really fragrant” after reading it. **

Not to mention, these photos are so exquisite that they are almost seamless, and even the neck lines on the neck can be reproduced 1:1, which made many users exclaim that “even my mother can’t fault it”.

In just one week after its launch, Miaoya Camera quickly became a hit. A large number of users flocked, which also made the waiting time for the photo to be generated up to 9 hours, and tens of thousands of people lined up in front. Roughly calculated, within a day, this small program can earn 100,000 yuan. According to this popularity estimate, it is not difficult to earn 1 million yuan in a week.

Although the waiting time is getting longer and longer, it still cannot stop the enthusiasm of users.

**After all, for 9.9 yuan, you can get a set of intensive “hippocampus” photos, and you can post them on Moments immediately. Who can be confused about this. **

In the pursuit of netizens, the first blockbuster in the AI era came into being, which came unexpectedly, but within reason.

** Of course, this is not the first AI image product in the AI era. **

A few days ago, Meitu Xiuxiu and Wink APP announced the launch of the “AI animation-modeling face” function. Users can upload photos to break through the dimension wall with one click and create their own exclusive animation modeling face.

In addition, at the “Meitu Image Festival” held by Meitu a month ago, CEO Wu Xinhong personally announced that Meitu will release 7 AI-related tools, including the visual model MiracleVision and Meitu Design Studio. Products, involving various AI image-related functions such as AI drawing, AI poster, AI product map, AI drawing, etc.

Soon after, Meitu also announced the “good news” brought by AIGC in the capital market.

In a July 18 announcement, Meitu stated that the company’s net profit increased significantly due to “rapid growth in revenue from the company’s artificial intelligence-generated content (AIGC)-driven VIP subscription business for images and video-related applications.”

However, there are many good news and continuous layout of the AI sector, but Meitu has never been able to usher in a “Miaoya Camera” of its own.

**Meitu AI, how effective is it? **

After All in AI, AI functions have taken a dominant position in Meitu Xiuxiu, and AI topic discussions have also become an important part of the Meitu community.

Apparently, for All in AI, Meitu seems determined to burn the boat.

But the concept must be implemented in the end. How effective it is depends on strength.

AI Blue Media Exchange first took the “AI drawing” function as the experimental object. In Meitu Xiuxiu’s “Wen Shengtu” dialog box, several keywords are input “a painting, lotus, light pink and light green tones, Zeng Chuangxing, close distance”, and the resulting effect is as follows:

It takes 3 seconds to generate four images with different styles and contents. On the whole, the outline of the details is relatively delicate, and the rendering effect on the portraits of the characters is quite normal.

At the same time, AI Blue Media Exchange also entered the same keyword in Wenxinyiyan and Bing AI, and the effect is as follows:

It also takes 3 seconds to generate, and Wenxinyiyan only provides one picture, and Bing AI is almost the same in terms of image detail and style diversity.

**On this item alone, Meitu has relatively obvious advantages. **

Let’s look at another set of comparisons.

In the “AI Elimination” function, AI Blue Media Exchange chose to eliminate the far-end passers-by. After 5 seconds, the effect given is as follows:

And the effect given by Xingtu APP, which also has the function of “passer-by elimination” is as follows:

Overall, the effect is not much different, but you can’t find it when you look closely. Meitu’s AI erasing function can choose to erase the area, but the effect on the original image is not preserved after erasing. For example, in this picture, the far-end The road map is erased, and the overall length of the street is shortened by AI;

Although the passerby elimination function of Xingtu cannot select the elimination area, there is generally no problem in judging the definition of “passerby”, and the effect of preserving the far-end images is better.

** In this wave, the elimination effects of the two methods are equal, but in terms of cost performance, Meitu needs to pay to become a VIP to be able to remove the watermark, while the wake-up image can be generated directly for free. **

In addition, there are other types of AI mapping functions in the Meitu app. For example, upload a selfie of a character to change the two-dimensional manga style with one click, “make a picture”, “make avatar”, “make a graffiti” into a picture, and “colorize a line draft”, etc.

In order to fully experience these functions, you must recharge to become a VIP member, and the membership price ranges from 12 yuan/month to 98 yuan/year.

At the same time, in Meitu’s community “gallery”, content such as AI beauties, AI photography, AI two-dimensional elements, etc. are almost full of pages. Although most users commented on the content, “the AI effect is really good-looking”, “can you exchange keywords”, but some users said, “although some of Meitu Xiuxiu’s AI functions are powerful enough, but some The function is the camera filter in the past, but now it is covered with a layer of AI, in the final analysis, it is just a concept”.

I have to worry about the beautiful picture

It is not difficult to find from personal experience and user comments: Meitu has a certain advantage in AI mapping, but the price is a hurdle, and the concept of being too close to AI has also given some users a bad experience.

But it is undeniable that AI has become and must become the core strategy of Meitu.

In order to alleviate users’ natural sensitivity and resistance to prices, Meitu has also begun to turn to B-side. In the “Design Room” page, the AI product map, AI poster and other functions are integrated, and the label “one-stop toolkit for merchants” is attached.

** In this regard, the capital market seems to pay extra. **

Just when Meitu released seven AI tools on July 18, the stock price of Meitu soared 21.28% that day. Xueqiu shows that in the past year, the stock price of Meitu has shown an overall upward trend, from 0.8 yuan per share to 3.5 yuan, with a total market value of 12.762 billion yuan. For this reason, some investors commented: “Buy from 2.60 to 0.65, and then buy from 0.65 to 3.55, insist on buying, without it, believe in yourself, believe in Meitu”, “Meitu AI is really serious about China aesthetics”.

In fact, for a long time, Meitu seems to be going further and further in the direction of “not doing business properly”, whether it is carrying out lending business or investing in speculation.

However, the arrival of the AI wave is like reining in a precipice and sending charcoal in a timely manner, giving Meitu the possibility of being trusted by the capital market again. In particular, Wu Xinhong also specifically mentioned that Meitu’s positioning should be returned to a tool—a transformative AI productivity tool, which is obviously more in line with the capital market’s expectations for products.

Even so, AIGC’s outlet is always a knife hanging over Meitu’s head.

Just like at the end of last year, AIGC was blowing out like a tide, which made many AIGC start-up companies show their edge. For example, Midjourney, a text-to-image product, is also the company that created the famous “Space Opera House” that is enough to be recorded in the history of AIGC. At the same time, this product can also generate exquisite and realistic portrait pictures by giving it detailed camera parameters.

**The effect is enough to replace the camera and photographer, not to mention the retouching software. **

It was also at this time that some people worried about the prospects of Meitu.

Such concerns are not groundless. To this day, Midjourney is still doing version upgrades. The new version V5 last month shows that the speed of generating images will be doubled, and the price will be cut in half.

In order to deal with such a menacing opponent, Meitu has been expanding its territory in the AI field almost every week in the past six months, involving image and video fields, covering people from B-end to C-end.

**This means that Meitu is anxious. **

Even so, in the context of continuous improvement in productivity, the first hit in the era of AI images still does not belong to Meitu. This situation is not a small blow to Meitu and investors who have expressed confidence in it.

Obviously, what Meitu needs to consider now is how to turn anxiety into hard power, instead of indulging in AI bubbles and getting deeper and deeper.