Midjourney Founder on the Future: Images Are Just the First Step—AI Will Disrupt Learning, Creativity, and Organizations
TechFlow Selected TechFlow Selected
Midjourney Founder on the Future: Images Are Just the First Step—AI Will Disrupt Learning, Creativity, and Organizations
11 people who changed the world.
Navigating Web3 tides with focused insightsText: Founder Park
Midjourney is a magical company—11 people changing the world, creating a great product.
It's destined to become a legendary tale of the early Pre-AGI era.
"I never wanted to start a company. I wanted to have a home."
At the BAAI Conference, we spoke with David Holz, founder of Midjourney.
Midjourney is currently the hottest image generation engine, maintaining an absolute lead in diverse style outputs despite fierce competition from OpenAI’s DALL·E 2 and open-source models like Stable Diffusion.
Below is the full transcript, edited by Founder Park.
Zhang Peng: Welcome David. I think this might be your first time communicating with Chinese fans. You can take this opportunity to say hello to them.
David Holz: Hi everyone! Midjourney may not have officially entered China yet, but if you're already using it, I hope you’re all having a great experience.
Entrepreneurial Motivation: Unleashing Human Imagination
Zhang Peng: Over the past 20 years, I've met many entrepreneurs from around the world. I’ve noticed they share something in common—they all have strong inner drives that push them to create something out of nothing.
I’d like to know, what was your driving force when you started Midjourney? What were you longing for at that moment?
David Holz: I never intended to start a company. I just wanted a "home".
I hope that over the next 10 or 20 years, I can build here at Midjourney the things I truly care about and want to bring into the world.
I often think about all kinds of problems. Maybe I can't solve every one, but I can make attempts so that everyone becomes more capable of solving problems themselves.
So I try to think about how to solve things, how to create. I think it comes down to three parts. First, we must reflect: What do we want? What exactly is the problem?
Then we must imagine: Where are we heading? What possibilities exist?
Finally, we must coordinate with others, collaborate, and collectively realize what we’ve imagined.
I believe AI offers a huge opportunity to integrate these three aspects and build important infrastructure that makes us better at doing this. In a way, AI should help us reflect on ourselves, better envision future directions, and better connect and collaborate with each other.
We can accomplish these things together and merge them into a single framework. I believe this will transform how we create and solve problems. That’s the big thing I want to do.
I think sometimes (what we do first) image generation might seem confusing, but in many ways, image generation is a well-established concept. Midjourney has become a collective super-imagination, with millions exploring the possibilities of this space together.
In the coming years, there will be far more visual and artistic exploration than the sum total of all previous history.
This won’t solve all our problems, but I see it as a test, an experiment. If we can achieve this kind of visual exploration together, we can do the same for other areas—anything that requires shared exploration and thinking. I believe similar approaches can solve those too.
So when I thought about where to begin, we had many ideas and built many prototypes. But then breakthroughs suddenly emerged in AI, especially in vision. We realized this was a unique opportunity to create something no one had ever tried before. That inspired us to give it a shot.
We believed that soon, everything would converge into something truly special. Now is just the beginning.
Zhang Peng: So image generation is just the first step. Your ultimate goal is to liberate human imagination. Is that what drew you to create Midjourney?
David Holz: I really love imaginative things. I also wish for a world with more creativity.
Seeing wild new ideas every day is incredibly exciting.
Serial Entrepreneurship: How to Build Better Startups?
Zhang Peng: Many people first knew you not through Midjourney, but through Leap Motion (David’s previous startup).
I strongly feel there might be some connection between the two projects—for you, whether in motivation or mindset. I'm not sure if I'm interpreting this correctly. But my question is: Did your experience at Leap Motion help you launch Midjourney?
David Holz: In many ways, I learned a lot from Leap Motion—how to build large research teams, how to tackle problems no one has solved before, how to deeply consider how people interact with technology.
It wasn’t just about speed, cost, and scale, but about how we collaborate and move forward together. This was crucial—it was one of Leap Motion’s core philosophies.
There are similarities—like Leap Motion, Midjourney grew rapidly and captured everyone’s imagination. I think Midjourney is similar in that sense.
But there are real differences too. At Leap Motion, we spent a lot of time building the early ecosystem.
In Silicon Valley, there was a belief back then that you should first build an ecosystem, like the iPhone. But what we’ve collectively learned is that maybe you should first build a truly great product that people love. Once you have that, you can start building the ecosystem.
For many entrepreneurs, focusing on ecosystem-building early on can be distracting, and I was no exception. So one key difference with Midjourney is that we focused on building something truly usable by everyone.
I used to judge my ability by asking, “Can I do it myself?” My father was a surgeon—he could. We have dexterous hands.
Now I think more: Would a truck driver enjoy using Midjourney? The fact that many truck drivers are using it blows me away—it means we’re doing something right.
Rethinking Knowledge: Historical Knowledge as Creative Power
Zhang Peng: That’s interesting. We usually say “ideas are cheap, show me the code.” But now, ideas seem to be the only thing that matters.
As long as you can express your idea through a series of excellent prompts, AI can help you realize it. So are the definitions of learning and creation changing? What do you think?
David Holz: One interesting thing is that when you give people more time to create, they naturally become more interested in learning itself.
For example, there’s a popular art style in the U.S. called Art Deco. I never cared about it—until one day, when I could generate Art Deco works via prompts, I suddenly became deeply curious and wanted to learn its history.
I find this fascinating—when history becomes something you can immediately use and create with, you naturally become more interested in it.
If the user interface becomes good enough that AI feels like an extension of our minds—if AI becomes part of our body and thought, and is deeply tied to history—then we too become more connected to history. That’s amazing.
When we ask users what they want most, the top responses are learning materials—not just how to use the tool, but about art, history, camera lenses, lighting—everything relevant to creation. They want to understand and master all knowledge and concepts available for creation.
Previously, knowledge was just the past. Now, knowledge becomes a creative force.
Knowledge now has immediate power. People are eager to gain more knowledge. That’s incredible.
Facing Anxiety: Openly Reflecting on Our Capabilities
Zhang Peng: In China, your users come from diverse backgrounds, with varying levels of artistic ability. Users like me are still struggling to make better images. But I’ve heard many artists and designers worry they’ll be replaced by AI.
To ensure everyone can benefit from AI instead of feeling confused or marginalized, what can Midjourney do?
David Holz: For those who haven’t started their artistic journey, Midjourney offers a unique opportunity.
You start asking: What do I like? What’s my aesthetic? What do I find beautiful?
Professional artists might spend decades developing this. Now anyone can do it. We’ve found that deep reflection is almost like art therapy—people think about their lives, challenges, what good or bad might happen. It’s very meaningful.
I think most users have these personal experiences—it’s not competitive or commercial. Most people use it for fun. They don’t even share the images they generate.
But in professional fields, it becomes an amplifier of creativity. Now you can create comics, films, or video games.
If you’re an ordinary person, thinking about beauty for the first time in your life, artists may now be thinking about creating entire worlds and universes—something they could never do before.
So it expands the boundaries of everyone’s creativity.
Ordinary people gain more capability, professionals gain more too. It’s hard to realize this because current interfaces are simple. But they can become more complex, powerful, and feature-rich. Those features will come later.
The Charm of AIGC: Is the Charm of Art Itself
Zhang Peng: Perhaps AI isn’t just for processing images—it can help people handle more complex tasks. Because in the past, creativity involved many factors: the desire to create, the ideas themselves, and the ability to execute and turn them into reality.
But now, if AI can free up creative desire, great ideas, and complex processes, it can assist you. That might be the power AI offers—not just images. Right?
David Holz: This is a very personal process, involving reflection—we didn’t realize this at first.
At first, I looked at these (user-generated) images and didn’t understand what people were trying to do.
Someone generated an image of a dog in heaven.
I asked why. He said his dog had just passed away.
I felt sad. He said the image made him feel better.
People use it almost like art therapy. They reflect on things, on their lives, trying to figure out who they are. I think this is deeply personal and profoundly important—something most people in human history never had the chance to do. Only a few ever did.
I think it’s beautiful that everyone can now do this.
Zhang Peng: Yes, Midjourney isn’t just a tool in our workflow—it’s become a new element in our lives.
The Legend of 11: More Will Emerge in the Next Five Years
Zhang Peng: Part of Midjourney’s magic lies in delivering magical technology to 10 million people. I know your team has only 20 people—just 11 months ago. No sales team, no marketing team, and some are even students.
I’m curious—is this a new model for startups in the AI era?
David Holz: I think so, yes.
We were among the first. We also had advantages—I have experience building teams, a decent reputation, access to computing power, and we started early.
Many typical startups lack these.
But in a few years, as people learn how to build research teams, as compute pressure eases, and as people better understand how to build great AI products, I believe we’ll see many such companies. We might be the first, but in five years, this could be commonplace.
General vs. Vertical: What Will Midjourney Become in the AGI Era?
Zhang Peng: Do you worry that what Midjourney does today might one day be absorbed by advancing AGI capabilities? What might happen?
What’s the future of vertical products like Midjourney?
David Holz: I don’t know. It’s a big mystery.
One possibility is that we might collaborate with other labs—we make the eyes, they make the ears, we each build different parts, then combine them. That could happen.
We create imagination, they build language models, and we integrate them. So we work on this together.
Another possibility: There may be general-purpose AGIs good at everything, but specialization will still exist. The world still has many specialists.
I also think there will be many human-AI interface challenges—not just making beautiful pictures, but helping people explore who they are and what they want.
This isn’t an AGI problem—it’s more about how to interact with humans.
The best interface might not just be language. AGI might ask questions, but actually, interfaces will show you many images, trying to understand people in different ways. It’ll be a new art, a new theory of understanding humans—AI helping people figure out what they truly want.
We haven’t thought about this much yet. AGI can do anything, but I believe specialization will still thrive in most future scenarios.
Power of Community: Exploring AI Copilot Together
Zhang Peng: Do you think Midjourney will always remain a vertical product, or could it evolve into something more general?
You’ve said your mission is to solve the problem of human conscious imagination—not just images.
David Holz: There’s something fascinating—we’re building a community, figuring out how groups can explore vast spaces together.
We’re trying to create this—in the next one or two years, Midjourney might become the place for nearly all visual exploration worldwide. So I think there’s something truly exciting here.
Even if we never do anything else, if we’re simply the global hub for visual exploration, I believe that’s valuable. But I also think visual exploration principles can apply to other domains.
Creating a super-mind of millions working together to solve problems is fascinating. I don’t think any single AGI can replace millions of people, but I do believe AI will participate—perhaps millions of people and millions of AIs working together. I think there will be many agents and participants, each with different perspectives, which helps deepen understanding.
Zhang Peng: Many so-called AI practitioners worry their products will be replaced by newer technologies.
Do you have the same fear—that what Midjourney does today might one day be superseded by evolving AGI?
What might the future look like for vertical products like Midjourney?
David Holz: Finding valuable problems worth solving, staying focused while remaining open-minded—that’s one side.
On the other hand, whatever AGI technologies or products emerge will still coexist in our society. We can keep using AI to create new things, and we can define how AI collaborates with us in creation.
That’s exactly what we’re doing—enabling collective exploration and co-creation.
Right now we’re doing it visually. In some ways, it’s primitive, but in others, it might be the best test—because you can see everything.
You can truly see it. If we did this with poetry, it’d be equally interesting—but harder to grasp. Or if we used engines or other sciences, you couldn’t easily find a million scientists, but finding a million image-makers is easy.
So this is really about imagining what the collective mind of AI and humanity should look like. Today, it’s mostly humans using AI to create. But in the future, AI and humans might think and create together—and that’ll be fascinating. Maybe someday, we won’t even be able to tell whether an idea came from a human or from AI.
Excessive Competitiveness: A Flaw in Human Nature
Zhang Peng: Of course, there’s growing concern and anxiety about AI safety and competition across various levels. Meanwhile, scaling and improving current large models remains challenging. How do you view the future development of AI?
How do you balance different viewpoints, even conflicting government interests?
David Holz: First, I think a flaw in human nature might be excessive competitiveness.
The truth is, there are countless opportunities and challenges in the world—so many problems mean so many opportunities. We face so many chances that we don’t even need to compete for them.
So for me, I don’t think about competing with anyone. There’s so much to do in the world—I just focus on creating what I believe is beautiful and astonishing.
When I was in China, I noticed how everyone collaborates closely. In a way, this resembles the principle of artificial intelligence.
Because it’s also highly communal—it learns from us, then feeds that ability back to us. It’s deeply community-oriented and collaborative.
I think competition isn’t always good. Sometimes even the desire “to be the best in a field” contradicts the essence of technology itself.
Moreover, we can learn much from Eastern culture. Where I am, people worry whether technology and AI are trustworthy enough. Yet elsewhere, people hope the world becomes smarter, so AI can help us create and solve more problems.
I noticed in China, people find you cool for doing creative, geeky things—and they want to be cool like you.
Now that everyone knows how powerful AI is, we can work together to provide infrastructure for everyone, building our living environment collectively.
Ten Years From Now: A Magical Future
Zhang Peng: One final question—if we look ahead ten years into the digital world, what changes do you think are inevitable in today’s human world?
David Holz: I believe in ten years, one person’s creativity might surpass that of an entire group today. When they come together, they’ll achieve things we can’t even imagine now.
These forces will create a better world.
Shared visions of a better future will bind people together, so we no longer fear the future but look forward to it with excitement.
A magical future surely awaits us—one filled with beauty. Approaching it with that mindset is the best way to realize that beautiful world. Of course, there are things I can’t fully grasp or foresee, but for the next decade, I think that’s how it should feel.
Zhang Peng: I agree with you.
Indeed, many wonderful things await us. As we move forward, we must also consider how to prevent the bad ones. This conversation has been incredibly insightful. Thank you so much for sharing!
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
