I’m going to be honest. I’ve spent over $25,000 on GPUs, cloud instances, and subscriptions to just about every AI image and video generator out there. Some people might call that a crazy amount of money to spend on a “hobby.” They might be right. But I didn’t just play around with prompts; I went deep. I wanted to understand the mechanics, the art, and the future of this new medium. And I can tell you that most of what you read online about AI art is surface-level nonsense.
It’s not about finding the “magic” prompt. It’s not about which model is “best.” It’s about composition. It’s about understanding how to translate an idea into a visually compelling image, using AI as a tool. And that’s a skill that costs time and money to develop. I’m writing this to save you some of both.
After all that investment, I’ve boiled down my experience into three critical lessons. These aren’t the obvious tips you’ll find in a YouTube tutorial. These are the hard-won insights that made the biggest difference in the quality of my work.
Lesson 1: Stop Obsessing Over Prompts, Start Studying Cinematography
When I first started, I was like everyone else. I spent hours on prompt engineering websites, copying and pasting long, convoluted prompts, hoping for a miracle. I’d get some cool-looking images, but they were inconsistent. I couldn’t reliably create what I had in my head. It was frustrating.
My breakthrough came when I stopped thinking like a prompt engineer and started thinking like a cinematographer. I have always been a movie buff, even before I sold MovieLaLa to Gfycat. I started to re-watch some of my favorite movies, but this time, I was paying attention to the composition of each shot. How was the frame balanced? Where was the light coming from? What kind of lens was used?
I started to incorporate cinematography terms into my prompts. Instead of just saying “a woman standing in a field,” I would write something like: “Wide-angle shot of a woman in a field, golden hour lighting, shallow depth of field, inspired by the cinematography of Emmanuel Lubezki.”
The difference was night and day. The images were more dynamic, more emotional, and more aligned with my creative vision. I realized that the AI models are trained on a massive dataset of images, including movie stills. They understand the language of cinematography, even if they don’t “understand” it in the human sense.
Here’s a practical tip: create a “shot list” for your AI art projects, just like a filmmaker would. Think about the different camera angles, lighting setups, and compositions you want to use. This will give you a clear roadmap and help you create a cohesive series of images.
Lesson 2: The “Perfect” Model Doesn’t Exist. Master a Workflow Instead.
I’ve seen so many debates online about which AI art generator is the best. Is it Midjourney? Stable Diffusion? Sora? The truth is, they all have their strengths and weaknesses. Midjourney is great for artistic flair, Stable Diffusion offers incredible control, and Sora is pushing the boundaries of video generation.
Instead of trying to find the “perfect” model, I’ve found it’s much more effective to develop a workflow that combines the strengths of different tools. For example, I might start by generating a base image in Midjourney, then use Stable Diffusion’s inpainting and outpainting features to refine the details and expand the canvas. I might even take the image into Photoshop for some final color correction and compositing.
This is where my experience as an angel investor in companies like Anthropic, OpenAI, Scale AI, and Hugging Face gives me a unique perspective. I see how these models are being developed and where the technology is heading. And I can tell you that the future is not about a single, monolithic AI that does everything. It’s about a constellation of specialized AIs that work together.
As an artist, your job is to be the conductor of this AI orchestra. You need to understand the capabilities of each instrument and how to combine them to create a beautiful symphony.
Here’s another practical tip: don’t be afraid to get your hands dirty with the technical side of things. Learn how to run Stable Diffusion locally on your own machine. Experiment with different settings and parameters. The more you understand how these models work under the hood, the more control you’ll have over the final output.
Lesson 3: Your Taste is Your Most Valuable Asset
This is the most important lesson of all. In a world where anyone can generate a technically perfect image with a few lines of text, the only thing that will set you apart is your taste. Your unique perspective, your creative vision, your ability to curate and combine ideas in a way that no one else can.
I’ve seen so many people get discouraged because they feel like they’re not “creative” enough to be an AI artist. They think that because the AI is doing most of the work, their own contribution is meaningless. But that’s completely backward.
The AI is a tool. It’s a very powerful tool, but it’s still just a tool. It can generate a million different images, but it can’t tell you which one is good. It can’t tell you which one is meaningful. It can’t tell you which one will resonate with people.
That’s your job. Your taste is the filter through which you process the infinite possibilities of AI art. It’s the compass that guides you toward a final image that is not just technically impressive, but also emotionally resonant.
So how do you develop your taste? You consume a lot of art. You study the masters. You go to museums. You read books. You watch movies. You listen to music. You expose yourself to a wide range of creative influences. And then you practice. You create, you experiment, you fail, and you learn from your mistakes.
I wrote a book called “Becoming Top 1%,” and the principles in that book apply here as well. It’s about deliberate practice, continuous learning, and developing a unique point of view. The same things that make a successful entrepreneur also make a successful artist.
The Road Ahead
AI art is not a fad. It’s a fundamental shift in how we create and consume visual media. It’s still early days, and there are a lot of unanswered questions. But one thing is for sure: the artists who succeed in this new paradigm will be the ones who combine technical skill with a strong creative vision.
So, my advice to you is this: don’t get bogged down in the hype. Don’t worry about which model is the “best.” Instead, focus on developing your skills and your taste. Learn the principles of art and cinematography. Master a workflow that works for you. And most importantly, have a strong opinion. In the age of AI, your humanity is your greatest strength.
Frequently Asked Questions
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.
How has this view evolved over time?
My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.
What's the most common pushback you get on this?
People often push back by citing exceptions or edge cases. And they're usually right that exceptions exist. But building a strategy around exceptions rather than patterns is a losing game for most founders.