"It's happening fast (the development speed is too fast)" Musk expressed this emotion on social platforms on February 12. He was talking about the ByteDance video model Seedance 2.0. This latest video generation model is no longer a simple technological event in the Internet circle, but a cognitive earthquake sweeping the entire chain of the film and television industry.

In the same week, director Jia Zhangke publicly stated that he would use Seedance 2.0 to produce short films; the film and television hurricane Tim called "horror" six times in a 9-minute measured video; the capital market took notice, and the net inflow of the main funds in the film and television media sector exceeded 3 billion yuan in a single day.

Feng Ji, the producer of "Black Myth: Wukong", posted a long post late at night, calling it "the most powerful video generation model currently on the planet" and asserting that "AIGC's childhood is over." Stephen Chow's agent Chen Zhenyu publicly questioned the infringement issue. The capital market also took note of the trend. On February 10, the A-share film and television sector broke out collectively, with the film and television theater sector rising 11.31% in a single day.

How could an AI tool stir up the entire film and television industry?

The answer is hidden in a number: the cost of a 5-second special effects shot dropped from 30,000 yuan to 3 yuan. More importantly, this is no longer a matter of "drawing cards" dozens of times to try your luck, but one generation and direct use.

When top-level visual effects directors take weeks of work and compress it into a two-minute production, the production logic of the film and television industry that has been running for nearly a hundred years is being quietly rewritten. But this change is far more complicated than "cost reduction and efficiency improvement" - some people see an unemployment crisis, some see golden opportunities, and some see deeper differences in routes.

Say goodbye to "drawing cards": From gamblers to directors, who are the winners and who are the losers?

Before the advent of Seedance 2.0, AI video creators had a self-deprecating nickname: Gambler.

In order to obtain facial features that do not collapse and a run that conforms to the laws of physics, it needs to be generated dozens or even hundreds of times. The industry calls this "drawing cards." Clothing textures suddenly flicker, backgrounds distort inexplicably, and the protagonist's face changes at every turn—these are commonplace. Previously, the industry average availability rate was only about 20%.

This uncertainty determines the fate of AI video in the old era: it can become a traffic code for Xiaohongshu bloggers to pursue, but it will never meet the delivery standards of the film and television industry.

Seedance 2.0 changes that. The actual measurement data of Jon Draper, founder of the British AI animation creative platform, is very illustrative: Compared with the model still in use a week ago, Seedance 2.0 has reduced the number of attempts required to generate usable images from 10 to 20 times to 1 to 5 times.

Chen Kun, founder of aipai.ai, also conducted actual testing. He used a prompt word of less than 100 words to generate a 15-second Journey to the West themed video in one go without any post-editing. His judgment was straightforward: this video has reached a level that can be commercially delivered in terms of lens logic, music, sound effects, and camera movement rationality. "If it is less than 15 seconds of content, it basically exceeds the level of the average director in the short drama industry."

The direct consequence of this “capacity flattening” is a complete reconstruction of the cost structure.Well-known visual effects director Yao Qi used this tool to produce a 2-minute science fiction short film "Homeward Bound" with an overall cost of only 330.6 yuan. Some people in the film and television industry estimate that the cost of generating a 5-second video can be reduced to 4.5 yuan to 9 yuan.

Feng Ji’s words are becoming reality: “The content field will surely experience unprecedented inflation.”

The first to feel the chill was at the executive level. According to Jon Draper's prediction, traditional animators, 3D animators, visual effects artists, camera crews, drone pilots, actors and audio engineers will be impacted; at the same time, the importance of art directors, producers, video editors/motion graphics designers will become increasingly prominent.

But overseas AI film and television producer Brett Stuart provides a calmer perspective. He said in an interview, "Seedance 2.0 does not make film production easier, it just makes it easier to get good-looking shots." He emphasized that the criterion for Hollywood's judgment is not that a shot looks good, but that hundreds or thousands of shots can withstand scrutiny and can be edited smoothly into a film.

Stuart believes that tools such as Seedance 2.0 will definitely subvert Hollywood, but they will not replace it suddenly. "The value of a professional director will be more reflected in taste, creation, performance shaping, shot scheduling and control of the final film effect - definitely not 'a person who can only type prompt words'. The director will not disappear. This work will become more clear: not to control the production resources, but to control the meaning of the work."

When the threshold of tools has been leveled, the core competitiveness of creators is drastically shifting from "how to shoot" to "what to shoot." Behind this is a cruel truth: when tools converge, the difference is no longer whether you can use them, but what you want to express.

From scissors to pen, who needs an editor?

If controllability is the first fortress that Seedance 2.0 conquers, then "multi-lens narrative" is a depth bomb it throws at the traditional film and television power structure.

All past AI video models have been stuck in "single shot generation" - it can present 15 seconds of beautiful footage, but cannot tell a complete story. The smallest unit of film and television narrative has never been the "shot", but the "edit". Montage, the last line of defense for human editors and directors, is beginning to loosen in the face of Seedance 2.0.

The experience of New York veteran documentary director Charles Curran is very representative. He used Seedance 2.0 to create a trailer for the live-action movie "Halo". The entire process took only 20 minutes and cost $60. This 1 minute and 24 second video showed movie-level special effects, precise audio and video synchronization, and professional multi-camera "shooting" effects, which made the director sigh, "This may really subvert Hollywood."

Brett Stuart of Staircase Studios AI studio has also released a number of animated short films produced with Seedance 2.0. He bluntly stated in his tweet: "Seedance 2.0 will completely change the future of film production. I am not exaggerating."

Stuart explained why he was so shocked: "The most amazing thing about Seedance 2.0 is the controllability, especially the multi-shot editing logic. The pictures seem to be carefully designed and scheduled, rather than moving for the sake of movement. Before this, AI could also generate moving shots, but they often lacked narrative."

The deeper changes are at the process level.In traditional film and television production, shooting and editing are strictly separated upstream and downstream: massive materials are produced on the set, and cut, reorganized, and given meaning in the post-production room. The logic of Seedance 2.0 is to generate and edit. At the moment it outputs a 60-second video, it has independently completed multi-angle camera scheduling, scene switching and emotional foreshadowing.

This fundamentally calls into question the role of the editor. From the perspective of industry applications, AI comics and dramas have become the most direct area to benefit. Huatai Securities predicts that the production cycle of a single episode of comics and other content may be compressed from the current 7 to 10 days to less than 3 days. In 2025, the size of China's comic drama market will be approximately 16.8 billion yuan, and it is expected to reach more than 24 billion yuan in 2026.

But the other side of the coin is that the living space for middle-level players is being drastically compressed. Huang Haonan, the founder of Soy Sauce Culture, said bluntly in his article: "When we are pursuing refinement and quality, Ji Meng 2.0 has set the table! Compared with the old animation companies such as Lingcage and Xuanji, can the so-called high-quality products be considered high-quality products? All high-quality products are built on large models."

The deeper industry speculation is: when AI begins to understand the lens language, what exactly are we witnessing? At the same time, the controversy also touches on a more fundamental question: When AI can generate images at high speed, is it "creating" or "mixing"?

Another big gap is performance accuracy. "Seedance 2.0 can generate stunning facial images, but the actors' delicate micro-expressions and precise control of timing, emotions and line nodes are still not stable and controllable enough. Moreover, when you add 'non-reality' elements to realistic scenes, the images tend to skew towards the CGI look and feel of games, closer to early special effects rather than the photo-realistic fusion of modern movie-levels."

The more powerful the tools, the “unique creativity of people” will become the only moat. When the rapid development of AI forces the growth of traditional industries, technology is doing an old-school thing: returning creation to the creator.

IP copyright owners usher in new business leverage

On February 9, Stephen Chow's agent Chen Zhenyu issued a series of questions on social platforms: "Are these infringements? I believe the creator should have made a profit, is the platform just letting it go?"

What triggered this question was the overwhelming popularity of "AI Stephen Chow" on short video platforms. With the computing power of Seedance 2.0, users' creativity has lost its boundaries: Mr. Xing's signature laughter, nonsensical expressions, and classic lines have been arbitrarily reproduced, and even inserted into absurd plots that have never happened.

This reveals another side of Seedance 2.0, which is both the Holy Grail of creative equality and a Pandora’s box of digital cloning.

Tim's actual test in the movie Hurricane Tim gave many people chills on their backs: He only uploaded a personal frontal photo without any sound files, and the model automatically generated an accent and tone that were highly similar to his own. He uploaded a photo of the front of the office building, and the AI ​​was able to "imagine" the actual layout of the other side of the building.

"It knows what's behind me, even if we don't tell it," Tim said in the video. He speculated, "This can basically confirm one thing, that is, Seedance 2.0 has trained a large amount of our company's videos."

This “licenseless cloning” capability quickly triggered regulatory brakes. On February 10, Jimeng officially suspended the ability to reference real-person materials, requiring users to complete real-person verification before they can create digital avatars, and restricting the generation and review of celebrity and well-known IP videos.

Sha Lei, a professor at the Artificial Intelligence Institute of Beihang University, pointed out a more fundamental issue: "The progress of AI will never stop due to controversy. How to find a balance between technological innovation, data compliance, and copyright protection is a common proposition for the global AI industry."

But past the fog of anxiety, a more profound business perspective is emerging: the digital assets of upstream IP parties are expected to be revalued.

While praising the technological breakthrough, Feng Ji also bluntly stated that this may bring about "the proliferation of fake videos and a crisis of trust", and said that "realistic fake videos will become without barriers, and the existing intellectual property rights and censorship systems will face unprecedented impacts."

This is the scarcity paradox in the AI ​​era.When tools are universally accessible and the technical gap at the execution level is smoothed out, “recognition” and “emotional accumulation” become the most luxurious moats. It is precisely because the market is flooded with a large number of "high imitations" of Stephen Chow that the "real Stephen Chow" IP is even more irreplaceable. When a massive amount of "high imitations of Stephen Chow" fill the information stream, users will still pay for the re-release of "God of Cookery"; when countless AI-generated Hollywood-level special effects shots can be obtained with one click, the sequel to "Avatar", which Cameron spent ten years polishing, is still a global event.

For copyright owners who own classic IP, they may usher in new leverage for commercial realization. Genuine authorization is no longer a simple image usage fee, but can rely on the enthusiasm of countless creators to achieve geometric amplification of IP assets under the fermentation of AI.

At present, online literary IP is concentrated in a few companies such as China Literature Group, Qimao, and Tomato Novel. Among them, Yuewen Group owns leading IPs such as "Celebrating More Than Years", "Sister-in-law", "Langya Bang", "Tomb Robbers Notes" and "Ghost Blowing the Lamp". China Literature quickly announced its access to the Seedance 2.0 model, which shows how seriously IP parties attach importance to this technological change.

Feng Ji said "AIGC's childhood is over." This sentence has two meanings.

The first level is ability: AI video finally gets rid of the clumsiness of babbling and can smoothly tell coherent audio-visual stories.

The second level is responsibility: childhood means trial and error can be tolerated, and recklessness can be forgiven. When a technology becomes "industrially available", it must face copyright litigation, ethical review and trust crisis.

It only takes two years from the launch of Sora in February 2024 to the market explosion of Seedance 2.0 in February 2026. Technology is evolving much faster than most people expect.

Brett Stuart's judgment may be closest to the truth: "Tools such as Seedance 2.0 will inevitably subvert Hollywood, but they will not replace it suddenly. The value of a professional director will be more reflected in taste, creation, performance shaping, shot scheduling and control of the final film effect - definitely not a 'person who can only type prompt words'. Directors will not disappear. This job will become more clear: not to control production resources, but to control the meaning of the work."

The film and television industry is standing at a fork in the road. Costs at the execution level are being infinitely compressed, and those jobs that rely on manpower and hours will be ruthlessly replaced. But at the same time,The value of IP and creativity is being infinitely magnified. When tools become readily available, what determines the quality of content will no longer be whether you can use software, but whether the idea of ​​the world in your mind is unique enough.

From DeepSeek to Seedance, China's AI has completed a leap from "catching up" to "leading"; and from Seedance to the future, what the entire society needs to face together is the journey from "technological confidence" to "civilization consciousness."

After all, it's never the stronger model that kills the game.

What kills the game is forgetting why we set out.