Yesterday, Dreamina 4.0 started its beta testing, and Nano Banana was released over a week ago. I plan to use both of these in my new films. These two mean that character consistency has been completely solved. I used to be an AI creator who was very competitive about character, scene, and style consistency. Back in the SD1.5 era, I was already training LoRA to maintain character consistency.

Can’t stop – this is probably the biggest daily feeling for AI creators. Working intensively from morning to night every day, trying to socialize a bit on content release days, then it’s on to the next piece. All tools are subscribed to on a monthly basis, and once paid for, they must be used completely, so there’s never a time to stop. Every day requires intense work – but is this intense work sustainable?

The story itself is the core of a film, and choices like non-linear narrative and visual artistry, without traditional storytelling, are all decisions made in the process of telling a story. But beyond the story, every design element that appears in the film is actually telling a story. As a director, everything needs to be designed.

When a person believes in something, they inevitably also have doubts. I have been researching and creating AI videos for nearly 2,000 hours and have met some friends in the industry. The development of all new things is not achieved overnight – they must go through twists and turns, and face opposition from those who resist them. Of course, there are also believers who push them forward vigorously.

Over the past three months, I finally had the opportunity to immerse myself wholeheartedly in creative work, whether in writing or producing AI videos. I completed the first book of “The Computational Universe” and rewrote the first half of the second book, allowing me to continue working on my novel. It felt like recovering from a long illness, letting go of many burdens.

When we discuss the soulfulness of a work, we usually refer to content that transcends life and death, time, and utilitarian purposes—topics that are eternal or related to life and death. I believe that the soulfulness of a work can be recognized at a glance by those who understand it, and they can also create such works.

I saw an article on Xiaohongshu (Little Red Book) arguing that AI is a form of mutual destruction, while its proponents exalt AI’s high efficiency and low barriers to entry. The article explains why AI often produces mediocre content. The author also believed that some people don’t seem to believe in science – they believe in technology, or more accurately, in the use of technology, or more specifically, in the use of weapons. In the end, they hoped for a great disaster to educate this anxiety-ridden society that prioritizes efficiency and productivity above all else.

Yesterday, I was chatting with a friend about the difficulties of AI technology infiltrating traditional film and television circles. Much of it stems from mindset issues, as well as the technical challenges of AI itself. For example, many people look down on ComfyUI because the Flux model currently lacks sufficient texture quality, and ComfyUI itself isn’t simple enough. Some even mock people who delve deep into AI technology, believing that aesthetics is the only thing that matters.

Yesterday, a group member asked: “Group admin, you used to be a writer. How did you become so skilled at video creation? Could you briefly share your learning process and experience?” So I’m writing this blog post to discuss my learning journey and insights. When it comes to technology, I’ve always held a relatively radical view – I don’t believe that focusing on the technical aspects of AI is useless. AI is a combination of technology, art, and literature. Only when these three elements are balanced can great works emerge.

This is a question raised by a group member that resonates with me. I share similar doubts because current AI-generated realistic styles suffer from a strong uncanny valley effect, with Veo3 perhaps being an exception. In other words, audiences may immediately recognize AI-generated realistic works as artificial, breaking their immersion. They tend to be much more tolerant of animated works.