An AI short series based on Cixin Liu's work hit 100 million views, but he says there will be no more sci-fi novels in the future | A conversation with Stanley Chan

AI闹·March 4, 2026·0·0

Produced by | AI Nao

01.

Before talking to Chen Qiufan, I'd had this vague feeling — that after the arrival of the AI era, most science fiction writing had become obsolete.

Chen Qiufan, a man who reached the future early. As a former sci-fi writer, he was seen as the leading figure of the new generation after Cixin Liu, "China's William Gibson." As a former tech insider, he'd worked at Google and Baidu. In 2021, he co-authored AI 2041 with Kai-Fu Lee, predicting the full arrival of the AI era before most people had even noticed.

But Chen agrees that sci-fi has become obsolete. When technological acceleration outpaces imagination, he says, that's the moment science fiction writers get eliminated.

Chen no longer writes science fiction. He assembled a team and made an AI short drama, The Divine Brush, which just finished airing on Mango TV with over 100 million views during the Spring Festival holiday. Cixin Liu served as executive producer; Chen was head writer; the director, Leo Zhu Zhili, is a PhD student in AI film and television research at The Hong Kong Polytechnic University — a brand-new program where both professors and students are beginners.

He teaches in Hong Kong, working with young people who jokingly call themselves NPCs of the real world, studying creative writing. These young people spend far more time talking to AI than to humans. The endings of their stories are often suicide. When grading their work, he can't help asking: Can't you give readers even a little bit of hope?

The first thing he does every morning is open his computer to talk to AI. He's a bit addicted.

Chen still writes fiction, but he'd rather say he writes realism, because AI has completely erased the boundary between fiction and reality.

02.

AI has arrived. Will humans become obsolete?

This may be the most hidden yet universal emotion of our era.

In Chen's The Divine Brush, the protagonist receives a magic brush she can use only five times, capable of helping her complete any masterpiece. But after the fifth use, she is swallowed by the system, becoming redundant data.

The entire drama points to a meta-question: When AI can accelerate, replace, and ultimately completely eliminate the human creative process, what remains of the person?

Chen resists obsolescence. His battle, at its core, is this: a person who once wielded imagination as a weapon, after imagination has been overtaken by reality, finding a new position for himself.

But when I revisit our conversation, I realize all of it was still about obsolescence.

This dialogue, perhaps, is for all those who are becoming obsolete.

"In Conversation with Chen Qiufan"

You Think You're Creating

But You're Just Feeding Each Other with AI

AI Nao As a former science fiction writer, why make an AI short drama?

Chen Qiufan I rarely call myself a sci-fi writer now.

At least within my reading range, most science fiction has become obsolete. In the past, when you read sci-fi, its imagination was at least a few steps ahead of reality. But now there are very few works with genuine technological foresight. Their imaginings are neither fresh nor interesting — compared to reality, they feel outdated.

So I say I write realism, because AI has erased the boundary between fiction and reality.

Making The Divine Brush was an experiment. Last Spring Festival, when DeepSeek had just blown up, I wrote this story. Later I attended Mango TV's annual meeting, got talking with people in the industry, and the opportunity came up. I thought: could I assemble a team and make an AI short drama myself?

AI Nao You wanted to be the one who assembles the team, not just the creator.

Chen Qiufan In recent years I've judged various AIGC competitions and noticed something. There are many creators in China who use AI very well, but few have managed to make the business model work. Many might take on some local promotional videos, some commercials, various piecemeal projects. But for culture to become an industry, there must be an ecosystem: sustainable production, commercial returns, market cultivation, audience groups, and distribution channels. Without this, much AIGC content remains self-entertainment.

So what I wanted to do was this. Making the drama was actually secondary — I wanted to run an experiment to see if this could work.

Then I brought in Leo — director Zhu Zhili. He's also a student in my class, simultaneously pursuing a PhD in AI film and television research. We're both very interested in AI's future in the film and television industry. His program is brand new; many questions are only just being considered globally, including tool usage, how to integrate into existing film and television workflows, and so on.

Now being a teacher isn't just about teaching — you also have to solve employment problems for students. Assembling this team and bringing them in was also about finding paths for students.

AI Nao In this drama's specific production, how high was AI's penetration rate?

Chen Qiufan Ours was very high — 80-90% in core segments, from scriptwriting, scene generation, to post-production. The broadcast feedback was also very good. Under Mango VIP's paywall, it exceeded 100 million views, making it a small viral hit.

During our creators' livestream, there were 11,000 people online.

AI Nao Are you yourself a short drama user?

Chen Qiufan I'm not. On the writing team there's another student of mine, Qianyi. She's a post-2000s woman, a heavy user of short dramas, web fiction, and games — her understanding of this domain far exceeds mine. She really knows what users want. A story has countless ways to be told; who ultimately decides how to tell it? It should be the users.

We extracted the core concept from the original work and rebuilt everything else — characters, plot, background — all aimed at short drama's target users.

AI Nao Cixin Liu as executive producer, former sci-fi writer running the show, plus a post-2000s short drama user as writer, and an AI PhD student as director. Your team is quite of-the-moment.

Chen Qiufan Indeed. Teacher Liu Cixin's involvement, plus many AIGC super-creators who joined later. But honestly, this combination itself was part of the experiment — today's technological frontier isn't in universities or sci-fi circles, it's in industry, whether startups or major tech companies. How else can we keep up? Beginners might be more willing to try.

AI Nao I watched the first few episodes. The protagonist gets swallowed by the system after using the magic brush five times. Last month I talked to Zhang Chu, and he had a formulation very similar to The Divine Brush's core — he said AI brings something like an LSD world, canceling the process and giving you the hallucinatory climax directly. This pleasure without cost eventually backfires.

Chen Qiufan To some degree, yes. But what I've observed is more complex.

People around me using AI to assist creation, including myself, very easily enter a "flow state" — can't stop. Whether learning or outputting, AI creates an extremely smooth, frictionless information feedback loop. I call this inter-intelligence: between two intelligent agents, you output to it, it outputs to you, you output to it again, infinite cycle. It feels great.

This is deeper than the pleasure of passive consumption. Before, watching short dramas was pure feeding; now you're also actively outputting, and AI reinforces you. This is next-level pleasure. This process even carries an illusion of productive labor — but what's worth warning against is, you think you're creating, but you're actually just feeding each other with AI.

AI Nao The state you describe — smooth, frictionless, infinite cycle — sounds like an upgraded version of the information cocoon.

Chen Qiufan Right, one step beyond the information cocoon. So I sometimes deliberately create friction.

Recently I've been using agents to build a queue, having them play different roles: scientist, writer, philosopher, some responsible for divergent thinking, some for challenging, some for integrating. I enjoy watching them debate each other. The result doesn't matter; what matters is the debate itself. Like returning to the ancient Greek agora. From your own perspective you can only ever have one viewpoint; often you can't break out of it. AI can give me many different perspectives.

AI Nao So you're actually actively pulling yourself out of inter-intelligence, letting the AIs fight each other while you watch from the side.

Chen Qiufan Yes. At this point you're more like a project manager, breaking down tasks, assigning them to different models, finally integrating results. Your judgment matters a lot in this process.

The Singularity Is Probably in 2027

But When the Singularity Arrives, People Won't Know It's the Singularity

AI Nao Starting from 2025, the world has been accelerating. This year has only just begun, and already OpenClaw, Seedance 2.0, Elys — viral events haven't stopped. As a former sci-fi writer, do you have this feeling of living inside a past prophecy?

Chen Qiufan Yes and no. Science fiction writers have prophesied too many things; people only remember the ones that were right, nobody remembers the wrong ones — survivor bias.

I think what's more important isn't whether prophecies are accurate, it's this speed — you just heard about something, and it's already outdated. Before, sci-fi still had space to think slowly; now reality runs faster than imagination, and narrative is being forced to keep up.

AI Nao Do you think sci-fi writers can't keep up with technological progress, or that technology itself has become too fast to be captured by narrative?

Chen Qiufan Both. On one hand, the technological frontier has completely moved into industry, into major tech companies. If you're not in that context, you simply can't follow.

On the other hand, disciplines are becoming increasingly vertical and specialized — quantum, synthetic biology, neuroscience. Even if you follow, you might not understand. Much of it is interdisciplinary, with very high cognitive demands, and then you still have to transform it into perceivable stories.

So why is The Three-Body Problem a classic? Its value actually isn't at the technical level, but in sociology, in structural dynamics — the Dark Forest, dimensional reduction strikes. It became a cognitive tool that people use to understand the world and anticipate the future.

But when sci-fi is treated as a cognitive tool, it also gets narrowed.

AI Nao So is the sci-fi genre only viable when technological development isn't accelerating? Like wuxia had its era, romance had its era — sci-fi was humanity's advance imagination before technological upheaval arrived. But now that the upheaval is really here, people don't have that space for reflection anymore.

Chen Qiufan Right. Look at the era of Frankenstein — electricity and biology, readers could still understand with their limited cognition. But now if you tell ordinary people about latent space, vector space, that's underlying logic requiring advanced mathematics, not something understood at the metaphorical level.

I think we've now entered an era approaching the singularity. Maybe next year, 2027.

AI Nao What are the signs of the singularity's arrival?

Chen Qiufan Here's what's interesting. Before the singularity arrives, or even when it does, no one knows what it looks like. All our guesses are only verified after the fact. Very possibly it will have happened next year, but we won't feel it. And looking back ten, twenty years later, oh, that was the starting point.

And the starting point isn't necessarily a point — it might be a period, calculated in 100 or 1000 years — in which case, we're already inside it.

AI Nao How do you feel now — excited, afraid?

Chen Qiufan Very complex, a bit bipolar. On one hand, the first thing I want to do every day is open my computer to talk to AI, a bit addicted. On the other hand, vaguely uneasy — teaching literature and creative writing at university, have we really prepared to welcome this upheaval? But this isn't something we get to choose; we can only be swept along.

Last time I discussed this with Chuan Wang, he said why do you always feel you need to control, to adapt? Maybe you should just surrender, submit to this upheaval — that might be a kind of wu wei happiness. I thought about it, and it made sense too.

People Like Us Still Have a Way to Survive

AI Nao This year's viral AI products like Moltbook, OpenClaw — have you used them?

Chen Qiufan Not yet. I feel the risk is a bit high; I'm thinking I'll wait until everyone else has tried them first (note: conversation took place in early February 2026). Some friends have used them, and I heard their whole computers got rummaged through. But I also wrote a new story because of this, not yet published.

AI Nao What kind of setup?

Chen Qiufan It's about art. Have you seen Zima Blue? The episode from Love, Death & Robots season one. I translated the original novella.

It's about an artist who paints across the stars, using a very pure blue, at enormous scale. A journalist interviews him, wanting to know why he repeatedly depicts this blue, and finally discovers his previous life was a pool-cleaning robot. What he did his whole life was endlessly clean every tile in the pool. That blue was the color of the tiles.

My new story somewhat imitates this idea. But it's about the future of crypto art.

AI Nao This story is powerful — you're giving me chills. I'm thinking you could sell it to Xingmai Innovation, a pool robot company that's been very hot these past two years.

Chen Qiufan Haha, could do. So returning to that question — why are humans so obsessed with creating machines that think like people? Maybe ultimately we want to understand — as humans, why can we think this way? Why do we have self-awareness? Why can we experience the world as so-called subjects?

I'm now skeptical of the word "subject" itself. Many people think AI has no subjectivity, but doesn't subjectivity emerge precisely through interaction with objects? Without the other, humans wouldn't have subjectivity or not. This is Daoist thought. Maybe when we interact with AI to a certain degree, its subjectivity naturally emerges.

AI Nao You teach creative writing in Hong Kong. Your students belong to the AI-native generation — how is their perception of the world different from yours?

Chen Qiufan I just finished a round of classes with them. Several students wrote stories ending in suicide. I said, can't you give readers even a little hope? They didn't say anything.

For many things, they ask AI first instead of talking to real people first. They definitely spend more time talking to AI than to people. This generation grew up in a rather challenging era; I feel many of them accept the world as a game-like setup.

AI Nao The world is a game — what are they?

Chen Qiufan NPCs, not protagonists, background characters, pushed forward by plot development or by larger trends, or passed by the protagonist.

AI Nao After so much abstract talk, let me ask you a practical question — over the past five years, how much has the money you've earned through AI risen as a percentage of your income?

Chen Qiufan Haha, the exact ratio is hard to say, but it's definitely much higher. In 2021, Kai-Fu Lee and I published the English edition of AI 2041, the simplified Chinese edition in 2022, and since then over twenty language editions. Many commercial activities, including royalties, are related to this book — AI really is helping me earn money.

Many people's AI entrepreneurship started just these two or three years, but I was earlier. Writing has a very high input-output ratio, with very long compound interest — as long as you're still alive and people are still reading your books, it keeps existing, and only grows.

AI Nao Last question. Science fiction writers are already an endangered species, but if someone still wants to write, including yourself, how should they write?

Chen Qiufan I can only speak for myself. Before, I very much avoided writing about personal experience; now I instead want to return to that. I might dig deeper into my own experiences and feelings, because those are things AI can't replace, can't generate. This way I might become a nonfiction writer? (laughs) But in the future, you won't be able to tell what's fiction and what's nonfiction anymore. What you think is fiction might actually be real.

I think the sci-fi label limits a creative work, and also limits a person. It gives others tinted glasses, viewing you through a particular perspective. So what I write in the future may not be, or not only be, science fiction — perhaps hard to define, ultimately only definable as "stuff written by Chen Qiufan."

AI Nao So what is writing to you now?

Chen Qiufan This is also what I tell my students — don't think about becoming famous, getting published. Writing is for yourself, to better understand yourself. It's also training in how to use language to express what you think and feel; this skill will help you communicate with AI and with people in the future.

Writing can be healing. Traumas you're unaware of, your emotional patterns, can be discovered through writing. When you see them, you know how to handle them.

AI Nao Sounds like a way for humans to return to themselves.

Chen Qiufan Right. All technology now, the AI explosion, is making you accelerate, making you transform, dismantling structures. Only creation is where you must face yourself, have a dialogue with yourself. Tools come and go, constantly iterating; you'll always have better tools, more realistic textures. But the ability to tell stories, to connect with people — if you don't have it, you just don't have it.

So I think people like us probably still have a way to survive, including my students. Through creation, many people in the AI era can actually find even firmer footing.

Join the Community