Seedance 2.0-Schockwelle: Ein Kostensturz von E-Commerce und Spielen bis hin zu Film und Fernsehen

While OpenAI overseas seems to have hit the "pause button" on the AI video generation model Sora, Chinese tech giants are launching a counterattack in this field.

Recently, ByteDance's latest AI video generation model, Seedance 2.0, went online and quickly set the internet ablaze with its advantages such as multimodal input, automatic camera guidance, and consistency.

After a in - depth experience with the model, Feng Ji, the founder of Game Science, made a very weighty assessment: "The content area will inevitably experience unprecedented inflation."

Feng Ji's prediction is not unfounded fear.

This wave of change is rapidly spreading to industries such as e - commerce, gaming, video platforms, and film production: In the e - commerce sector, the technical barriers for low - end outsourcing companies and photo studios have completely fallen. In the gaming industry, the production cycle for concept validation and promotional materials is reduced to a minimum, and the competition is getting fiercer. Video platforms must further optimize their distribution logic to cope with the sudden increase in supply. And the traditional linear process of film production "shooting + editing" is being decimated by the industrial pipeline "prompt + generation".

An industrial restructuring in terms of benefits and substitution has already begun.

1

The Explosion of Video Production Capacity

Last year, the biggest problem with AI videos was deliverability.

Whether it was Sora, Runway, or the Chinese models Keling, and even Jimeng developed by ByteDance itself - all had such problems. Creators often found themselves in a kind of "lottery" to get a few seconds of non - faulty and consistent video after dozens of repetitions.

The key point of Seedance 2.0's breakthrough lies in transforming the "show" into a "deliverable narrative".

The breakthrough in key competencies is mainly manifested in three aspects:

First, multimodal input. According to tests by All - Weather Tech, newly registered premium users of Jimeng Seedance 2.0 can use it for only 1 yuan with automatic renewal. It accepts text, images, videos, and audio as reference materials. Almost all imaginable formats can be used for video creation.

Second, understanding of narratives and the ability of automatic camera guidance. Seedance 2.0 shows a "director - led" thinking. It not only understands complex narrative logics but also can automatically control camera language and perform camera movements such as panning, zooming, tilting, and rotating. The video is no longer a simple shift of static images but has a cinematic narrative logic.

Third, the consistency of images. According to tests by All - Weather Tech on several AI video generation applications in the market, problems such as distorted facial expressions during the movement of the main object and blurry background images occur repeatedly.

But from the demo videos, it can be seen that Seedance 2.0 maintains the consistency of faces and images during the movement of the main object, which enables a coherent narrative.

This means that AI video creation has evolved from a toy to a tool. This ability to transform video creation into a standardized industrial pipeline makes the statement "Everyone is a director" no longer an empty promise and will significantly reduce the cost of video production.

Feng Ji describes this change with the term "inflation".

"The production costs of ordinary videos can no longer be calculated according to the traditional logic of the film industry and are gradually approaching the marginal cost of computing power. The content area will inevitably experience unprecedented inflation, and the traditional organization and production process will be completely restructured. I'm sure that all those who have used the model will quickly understand that this prediction is not unfounded fear." said Feng Ji.

2

The First Wave of Impact

When the marginal cost of video production approaches zero, the business models based on the old cost structure will be most affected.

The four industries of e - commerce, gaming, video platforms, and film production may be the first to be affected by these changes.

The most direct impact is first seen in the e - commerce sector.

Product presentations, scenario demonstrations, and function explanation videos essentially do not depend on complex artistic narratives but on clear information transmission.

With the spread of Seedance 2.0, the barriers for merchants to acquire video presentation capabilities have completely fallen. Low - end video outsourcing companies and Taobao photo studios, which used to profit from "information differences" and "technical barriers", will face a cold spell. Video creation could shift from professional outsourcing services to the daily activities of merchants themselves.

Compared with e - commerce, the impact of AI video generation models on the gaming industry may still be relatively limited, but the revolution has already begun quietly.

The costs of presenting worldviews, validating concepts, and producing promotional materials for game videos are decreasing exponentially. More projects will be validated in earlier phases and also rejected in earlier phases.

An employee of a game company in Beijing told All - Weather Tech that the company has already started a small - scale test phase of Seedance 2.0 internally.

AI video generation models are also changing the distribution logic of video platforms.

For platforms like Douyin and Kuaishou, the videos generated by models like Seedance 2.0 mean a sudden increase in content supply. This forces the platforms to completely shift their core competence to "filtering and distribution". For example, the winner will be the platform whose algorithm can more precisely select valuable content from the large amount of AI - generated content and whose commercial implementation efficiency is higher.

In the film industry, Seedance 2.0's ability for multi - line narrative could reshape the production process.

In the past, the creation of a film followed a strict linear industrial process: First, a large amount of materials were shot, and then these were selected, assembled by editors in the post - production room, and the narrative logic was established.

But in the logic of Seedance 2.0, this boundary is becoming increasingly blurred.

In the future shooting process, there is a possibility that the sets will be inexpensively generated by AI models. The model itself understands camera guidance and narrative rhythm, and when generating the video, the "editing work" is essentially already completed synchronously.

The AI no longer just provides individual material shots but directly a "finished film" with a coherent spatio - temporal relationship.

This means that the time - consuming post - production step in traditional film production is at risk of being "decimated" by algorithms.

The future creative process may no longer be "shooting + editing" but "prompt + generation". The role of editors will change from "workers" to "command engineers" or "aesthetic reviewers".

Although the videos currently generated by Seedance 2.0 are not 100% perfect and still need improvements in logical details and images, these problems will no longer be obstacles in the near future, given the rapid technological development.

3

The "Protection Wall" of IP

Seedance 2.0's amazing "imitative ability" allows ordinary people to enjoy the pleasure of creativity, but it also brings unprecedented pressure to copyright holders.

Recently, numerous "two - part" or even "parody" clips from Zhou Xingchi's classic films have gone viral on short - video platforms.

With the computing power of AI video generation models, Zhou Xingchi's facial expressions, his distinctive laughter, and even his classic language style have been inexpensively imitated by many users, and many absurd scenarios that have never happened are even generated.

This has quickly caught the attention of Zhou Xingchi's team.

Chen Zhenyu, Zhou Xingchi's manager, publicly wrote: "I want to ask, is this an infringement of copyright (especially in recent days when there have been many spreads)? I believe that the creators are already making profits. And does a certain platform simply allow users to generate and publish such content?"

This question seems to point out the fear of copyright in the AI era, but from a commercial logic perspective, it actually proves the opposite: Top - tier IPs are extremely rare in the AI era.

In the future flood of AI - generated content, technology itself will no longer be a barrier, because everyone has access to the same Seedance 2.0 tool.

The real barriers still lie in the hands of IP owners.

Precisely because the market is full of "clones" of Zhou Xingchi, the "real" Zhou Xingchi IP becomes even more indispensable.

When the content supply not only becomes excessive but also "inflationary", users' time and attention will become incredibly expensive. Those that can immediately capture users' attention are still the classic IPs that have withstood the test of time and have strong emotional penetration.

In other words: Although AI lowers the barriers for production, it infinitely increases the value of "distinguishability".

For IP owners, the future still looks rosy. The IP assets accumulated over the years will not only be victims of copyright infringement but can achieve an exponential increase in commercial value through the licensing of original rights and the AI machinery in the hands of countless creatives.

From the release of OpenAI's Sora 1.0 in February 2024, the world's first AI video generation model that can generate videos up to 60 seconds long, to today's ability of ByteDance's Seedance 2.0 to generate 60 - second feature films with multimodal input and original audio narration, only two years have passed.

In an era of rapid technological development, all industries are at a crossroads: The costs at the execution level will be infinitely reduced, and jobs based on repetition, labor, and time consumption will be mercilessly replaced. At the same time, the value of IP and creativity will be infinitely magnified.

When tools are easily accessible, it will no longer be the ability to operate software that determines the quality of content, but people's imagination of how the world should look.

This article is from the WeChat account "All - Weather Tech" (ID: iawtmt). Author: Zheng Minfang, Editor: Song He. Published by 36Kr with permission.