
If you haven’t seen the DeepSeek Big Fat Fish and Miss Claude skits, then I don’t want to talk to you ( bushi
Foreigners like this who have never eaten fine rice bran are still rude at first.

But after just one stroll around Station B, I immediately changed my tune.

After a few days of immersive tasting, the painting style has become like this. . .

From rejection to curiosity, from not understanding to saying I want it, all it takes is a few high-quality AI videos.
There are also those who are better at doing things, like this one in “Prime”, where a romance drama started right next to the beach villa.
Just click on Glanze, its publishing platform, and you don’t have to worry about downloading dramas or anything like that.
Shichao chased and gnawed day and night, and then a violent comment came to his mind:
AI-generated videos and short plays may really become a brand new art in the future.
Hey, don’t rush to hit me with a brick. In fact, I refused at the beginning.
After all, this thing is famous for its brainless coolness. Forget about the industrial saccharine, it has an all-AI face that can even be used in the uncanny valley. There is a reason why it is said to be digital swill.
But after thinking about it more, I discovered that swilling water may actually be necessary to create good art.
Because when things like novels and movies first came out, they were about the same as AI short plays. Later, after struggling in the swill bucket for several years, I finally found some works that can be called art.
Take America at the end of the 19th century as an example. The most popular reading material at that time was “Pulp Ficiton”.
This thing features a brainless, rough and over-the-top character. The cover is either a muscular man or a blonde beauty. If it’s not strong enough, he can even insert an alien Squidward for you. The plot is completely unrestrained. It’s not afraid of being disrespectful, but it’s afraid of being unreliable.
Its popularity was mainly due to cheap wood pulp paper (Pulp) and high-speed rotary printing presses, which made the book printing business a low-priced business.
Once the cost is reduced, everything that can be printed will be gone. Readers enjoy reading, publishers make money, and pulp novels are a direct winner.
In addition to fast food and cool writing, what’s more important is that it supports a group of quite capable writers.
Like Lovecraft, who created the early world view of Cthulhu, veterans like Asimov and Heinlein gradually pushed hard-core science fiction to the altar.
Therefore, new technology will not dictate what content can be produced, but it will expand the content market.
As for how to create real art, it still depends on the creators to explore bit by bit.
The same goes for movies that are regarded as the seventh art. More than a hundred years ago, when the Lumiere brothers invented the movie camera, the industry was also quite confused.
The director set up a camera and shot whatever was in front of him, just like today’s surveillance system.
Therefore, most of the early movies were slices of life of tens of seconds, which could be collected when a train arrives at a station or at a factory gate, and played on a loop in amusement parks and vaudeville theaters for everyone to watch.
In order: “Train Arrives at the Station”, “Factory Gate”, “Gardener Watering”
It wasn’t until Méliès started playing around with stoppage replacements and set special effects, Bout tried to cut different shots together, and everyone started tinkering with scenery and narrative editing that the film discovered that it could use lenses to reorganize time, space, and emotions.
As these guys matured, they slowly took off the label of “moving photos” and developed their own artistic language.
So when I look back at the current AI video, many things become clear.
Video models such as Seedance and Kling have broken through the threshold, just like when wood pulp paper was popularized and the Lumiere brothers invented the movie machine.
You have to expand the plate first so that the good things can be slowly fished out from the swill.
As for the current mixed bag, we cannot blame the creators for their lack of aesthetics. After all, AI is a new machine that everyone has just touched, and it is basically still following the old methods of traditional film and television.
Slicing for a few seconds is fine, but when it comes to a large-scale production, new problems such as character consistency, scene connection, and camera movement control arise, and the old routine becomes unusable.
There are indeed many solutions on the market, but the effect of adaptation depends on taking it one step at a time. This set of saddles from AIGC has only just been put together. . .
For example, “Pai Zi” topped the popularity list of station B and has now been played by more than 20 million people.
When we interviewed the author DiDi_OK before, we learned that to create a work of this level, the idea in the mind is the first step, and the more torturous part is actually dealing with the model.
To put it bluntly, it’s like trying your luck to draw cards, round after round, 3,000 times in one shot.
Even though the TapNow he was using at the time was already the industry’s top AIGC creation tool, when he opened the canvas, it was still densely packed with nodes and connections. Those who didn’t know it would have thought they were entering an old computer room that was decades old. . .
Even our later classmates in the editorial department often complain that the traditional canvas mode is actually not very user-friendly, and most of the energy is spent on Lianlian.
Think about it, the same Workflow (Workflow) on the big language model has long since faded away, giving way to smoother Agent interaction. Like Coze, which was very popular before, it has lost its voice recently.
If AI videos remain at the stage of drawing cards in manual workshops, it will indeed be difficult to produce any popular art.
So what do the current AIGC creation tools look like? Have you followed Agent’s wave of evolution?
After a long time, we opened TapNow again.
Hey, don’t tell me, I just listened to people’s advice and had a full meal. After experiencing this interactive effect, I feel it’s a bit beyond expectations.
Especially for new things, they are updated sparsely, as if they are free of charge. In addition to making pictures and videos of the model, it can also create small animations rendered by pure code. Anyway, various Agents and Skills are used together, and it is quite AI Native to use.
Like the most commonly used canvas, it has been iterated to the Creative OS version 2.0. Say goodbye to the traditional Lianliankan, and a group of Agents are stationed on it to help you do chores.
You don’t even need to type. If you need it, just hold down the spacebar and open your mouth to press BB. After using it a few times, Shichao is already in the form of voice input. . .
In order to test the effect, we made a short short film from scratch and found that we no longer have to worry about chores and only focus on the final film, which is more efficient.
For example, I have seen a lot of second creations of Big Fat Fish recently, so I used it to find some inspiration.
Then it directly aroused a bunch of Agents to search the Internet, and soon found several AI anthropomorphic images.
Looking at Whale Girl, GPT Dragon Girl and Miss Claude, the inspiration came to me, and I directly generated three views to create a “Western AI Red Dead Redemption”.
Just open your mouth, and then you can arrange a few agents to create a cowboy version of the character, and then complete the scene as a reference.
After finalizing the script, the small theater can officially start filming.
The most exciting thing is this part of the shooting. Before, you had to connect and correspond to the assets yourself. Now you don’t have to worry about it at all. You can just let the Agent connect by itself.
Seedance2.5 comes with a comprehensive reference, so characters, scenes, props, etc. will all come out by themselves and connect the lines.
Since the script was about one minute long and could not be generated at once, it split the script into two parts and planned the storyboards to work on both sides.
The video was created in a short time, and the setting diagram was buried in front, so the character consistency was stable.
I also took a shot from the perspective of the coin, and the lens movement and zooming were very smooth, and it really felt good.
However, after just one operation, there were a lot of nodes on the canvas, which made my scalp numb just by looking at it.
Fortunately, the optimization is not bad, let the Agent take action, organize it in a few seconds, and then package it for everyone to taste.
That’s right, this thing can be seamlessly converted into a team space, and students from the next stage can be invited to come in and give pointers together.
And if you just got started and have no clue, it doesn’t matter. There are various lazy packages in the TapNow community. Many skills and templates have been optimized by big guys. You can use them by modifying them.
Especially since the official himself is also a trainer and will get you some hard stuff from time to time, like this one he got a few days ago. Cinemap。
We experienced it in depth and found that this little thing is really fun. It is basically a real-time studio with unlimited DIY.
I won’t go into the pre-made western towns, New York and London, etc. You can create any world view you want. The key is that it is still in the closed beta period, and there is no charge at all.
What else is there to say? Just push hard and the 3D sandbox world will start directly.
Shichao immediately decided to create a setting of Harry Potter stitching Chinese immortality, and asked Agent to shoot two scenes to test the water.
In the end, within a few minutes, it was actually set up on the set. By dragging the progress bar, you can see the white model figures flying around on the map.
The key is that in addition to letting Agent do the work, you can also do it yourself. Just scan the code with your phone to simulate the perspective of a drone and create an aerial shot.
Shichao also took advantage of the situation and used Porter to ride on his broom, pass through the mountain gate, fly over the waterfall, and circle around the main hall.
After recording, you can directly import the video to the canvas, and Agent will render it for you.
This time, Seedance2.5 is on the good side. It no longer has to endure Shichao’s left-brain-attacking-right-brain prompts. It points to where to hit, and the effect is very comfortable.
By the way, the video above was not manually spelled, it was made directly by pure AI.
Because TapNow also integrates a Monta editor, the logic is the same as that of Cinemap. Import the video from the canvas to the AI editor, then hold down the spacebar and BB, and the video will be assembled. . .
Of course, although the whole process is smooth, to be honest, the price of generating videos is still a pain in the ass.
Fortunately, TapNow is still running a promotion. I took a look and saw that the quarterly package is 40% off, and the yearly package is half off.
This is pretty standard, but what’s even more exciting is the free prostitution period offered by high-end tiers. Seedance 2.5 can be used unlimitedly in the first few days after opening, so you must take advantage of this.
After finishing the process, I stared at the few little octopuses wandering around on the screen, which was quite emotional.
Not to mention that in the past, dozens of people carried millions of dollars of equipment and were busy setting up scenes and editing to capture the images. Now one person can do it.
The magic of voice interaction is greater than I imagined. Sometimes using your mouth is really more convenient than typing on the keyboard.
Such a low threshold for creation is really eye-watering.
I couldn’t help itching my hands, so I rubbed them again:
We don’t know if this thing can take over Hollywood, but the threshold is so low that a vigorous swill war is inevitable.
Therefore, instead of worrying about whether there is gold in the swill, Shichao chose to take the initiative this time and start enjoying it directly.
After all, this tool is in everyone’s hands, so we don’t have to wait for a big director’s talk.
AIGC’s new art form may be hidden in your random thoughts.
Maybe in the middle of the night one day, you open the Creative OS updated to version 42.0, and among the thousands of nodes, a genius idea comes to you.
Then the door to this new world might be knocked open. . .
Related Reading
- Anthropic engineer’s warning about Claude Code explains why tech outages could get much worse2026-11-22
- Original Unexpectedly, just one day after the APEC summit, China became famous around the world because of its three “specialties”2026-11-21
- Original Anthropic goes public and creates wealth: 96% of employees became millionaires overnight2026-11-20
- The beauties and handsome guys from UB.com want to enter your home and wait until they arrive.2026-11-20
- Original: Tencent’s shareholders will suffer again2026-11-20