How to Use Midjourney to Create Results That Got 10M+ Views

I’m going to show you how I used Midjourney to create images that have been seen by over ten million people. And no, it wasn’t by getting lucky with a viral prompt.

My name is Sahin Boydas. I’m a builder and an investor in Silicon Valley. I’ve been fortunate enough to have a couple of successful exits – you might have heard of RemoteTeam, which was acquired by Gusto, or MovieLaLa, which is now part of Gfycat. I’ve also been an early investor in some of the companies that are shaping the future of AI, like Anthropic, OpenAI, Scale AI, and Hugging Face. I’m not saying this to brag, but to give you some context. I’ve seen what it takes to build things that people love, and I’ve seen how technology can be a massive amplifier.

This isn’t going to be another one of those “10 amazing Midjourney prompts” articles. There are enough of those already. This is a strategic guide for people who want to use AI to create work that gets real-world results. I’m talking about results that you can measure – in views, in engagement, in revenue. I’m going to share the exact framework I’ve used to generate images that have gotten over 10 million views. It’s a framework that I’ve developed through countless hours of experimentation, and it’s the same one I share with the founders of my portfolio companies.

So, if you’re ready to move beyond generic, forgettable AI art and start creating work that makes an impact, you’re in the right place.

The Problem with “Good Enough” AI Art

Most people using AI image generators are stuck in a loop. They type in a basic prompt, get a decent-looking image, and call it a day. It’s “good enough.” But “good enough” doesn’t get you ten million views. “Good enough” doesn’t build a brand. It just adds to the noise.

Back when we were building MovieLaLa, we lived and breathed visual content. We had to create compelling images for thousands of movies, and we had to do it on a startup budget. We learned that the right image could make the difference between a user clicking or scrolling past. The image was the hook. It had to tell a story in a fraction of a second. We didn’t have Midjourney back then, so we had to get creative with stock photos and graphic designers. It was a grind.

Today, you have a tool that can create literally any image you can imagine. It’s an insane advantage. But most people are squandering it. They’re treating it like a stock photo library on steroids, pulling out generic images that look like everything else on the internet. Your audience is getting smarter. They can spot lazy AI content from a mile away. It has a certain look, a certain soulless perfection. And they’re starting to ignore it.

If you want to stand out, you need to stop thinking like a prompt monkey and start thinking like a creative director. You need a system. You need a framework.

My Framework for 10M+ Views: The Art of Strategic Image Generation

This is the exact framework I use, and it’s what I teach the founders I invest in. It’s not about finding the “perfect” prompt. It’s about a systematic approach to creating images that serve a specific goal. I call it the Strategic Image Generation Framework.

It breaks down into four key stages:

1. Define the Core Objective: What is this image for? Is it to stop the scroll on social media? Is it to illustrate a complex point in a blog post? Is it for a landing page to drive conversions? You have to be brutally clear about the job to be done. An image designed for a tweet is very different from a hero image for a website.

2. Establish the Visual Language: This is about brand. What is your visual identity? Are you dark and moody, or bright and optimistic? Are you minimalist, or are you detailed and complex? You need to define a consistent visual style. This is how people start to recognize your work. When I was investing in Scale AI, we talked a lot about how to visually represent the complexity of data labeling. It couldn’t just be generic servers and code. It had to feel human, precise, and powerful.

3. Iterative Prompting & Curation: This is where the magic happens. It’s not about one-shot prompting. It’s a conversation with the AI. You start with a base prompt, and then you refine, remix, and regenerate. You might generate 20, 50, or even 100 variations of an image. The goal is to explore the creative space and then curate ruthlessly. You’re looking for the one image that doesn’t just look good, but that also nails the objective and the visual language.

4. Post-Production & Polish: No AI image is perfect out of the box. The final 10% is what separates the amateurs from the pros. This could be anything from a simple color correction in Photoshop to compositing multiple AI-generated elements together. It’s about adding that human touch, that final layer of polish that makes the image truly yours.

Let’s break down each of these stages.

Stage 1: Define the Core Objective

This sounds obvious, but it’s the step that 99% of people skip. They jump straight to thinking about what they want to see, not what they want to achieve. Before you write a single word of a prompt, you need to answer this question: What is the single most important thing I want this image to do?

• Stop the scroll? This means you need high contrast, a clear focal point, and an element of surprise.
• Explain a concept? This means you need clarity, simplicity, and a visual metaphor that makes sense instantly.
• Build a mood? This means you need to focus on color, lighting, and composition.
• Drive a click? This often means including a human face, showing a desirable outcome, or creating a sense of curiosity.

Let me give you a real-world example. A while back, I wanted to write a post about the future of AI in robotics. I had this idea of a robot that wasn’t cold and industrial, but one that was almost a part of nature. The objective wasn’t just to show a robot. It was to make people feel a sense of wonder and optimism about the future of AI. I wanted to create an image that felt more like a piece of art than a tech diagram.

My objective was clear: Create an image that makes people feel optimistic and curious about the intersection of AI and nature.

That single sentence guided every decision I made afterward. It told me I needed soft lighting, organic shapes, and a natural setting. It told me to avoid hard edges, metallic sheens, and sterile environments. Without that objective, I would have just prompted “a futuristic robot” and gotten something generic. Instead, I got an image that people connected with on an emotional level. That’s the power of a clear objective.

Before you even think about opening Midjourney, write down your objective. One sentence. Be specific. Be ruthless. If you can’t define it, you’re just playing around. Which is fine, but it won’t get you to 10 million views.

Stage 2: Establish the Visual Language

Once you know your objective, you need to define your style. This is your brand. It’s what makes your work recognizable. If you’re just starting out, you might not have a defined visual language yet. That’s okay. This is your chance to create one.

Think about the brands you admire. They all have a consistent visual style. Apple is clean, minimalist, and premium. Nike is dynamic, high-contrast, and aspirational. You need to do the same for your own work. Your visual language is a combination of several elements:

• Color Palette: Are you using a limited color palette? Are your colors saturated or desaturated? Warm or cool?
• Composition: Are your images centered and symmetrical, or do you use the rule of thirds? Are they busy or spacious?
• Lighting: Is your lighting dramatic and high-contrast, or soft and diffused? Is it natural or artificial?
• Medium: Does your work look like a photograph, a painting, a 3D render, or something else entirely? Midjourney can emulate almost any medium, so you have a huge amount of control here.

When I was working on the robot-in-nature concept, I knew I wanted a visual language that was a mix of sci-fi and naturalism. I wanted it to feel like a photograph from a future National Geographic. So, my visual language was:

• Color Palette: Earth tones, with one accent color for the robot’s glowing elements.
• Composition: Rule of thirds, with the robot off-center to create a sense of discovery.
• Lighting: Soft, diffused morning light, as if the robot was just discovered at dawn.
• Medium: Photorealistic, with a shallow depth of field to make the robot pop from the background.

This is the level of detail you need to be thinking at. Don’t just say “I want it to look cool.” Define why it will look cool. This is how you build a recognizable style. It’s also how you make your prompts more effective. Instead of just prompting for “a robot in a forest,” I could now prompt for “a photorealistic shot of a bio-mechanical robot in a misty forest at dawn, soft morning light, shallow depth of field, earth tones with a single glowing blue accent.” See the difference? One is a vague idea. The other is a creative direction.

If you’re serious about this, create a style guide for yourself. A simple document with your color palette, your go-to composition styles, and your preferred lighting and medium. It will save you hours of work and make your output infinitely more consistent.

Stage 3: Iterative Prompting & Curation

This is where the real work begins. It’s also where most people give up. They write a prompt, get a result they don’t like, and either tweak a word or two or abandon the idea altogether. That’s the wrong way to think about it. You need to think of it as a conversation. The AI is your creative partner, not a vending machine. You give it an idea, it shows you what it thinks you mean, and then you refine and redirect.

My process here is very similar to how I approach angel investing. I don’t just find one company and bet everything on it. I look at dozens of companies in a space. I make a lot of small bets. I see what gets traction. Then I double down on the winners. You should do the same with your prompts.

Here’s my process for iterative prompting:

1. The Base Prompt: Start with the detailed prompt you developed in the previous stage. This is your starting point. For my robot in the forest, it was: “a photorealistic shot of a bio-mechanical robot in a misty forest at dawn, soft morning light, shallow depth of field, earth tones with a single glowing blue accent.”

2. Generate Variations: I’ll take that prompt and generate at least 4-8 initial images. I’m not looking for a perfect image yet. I’m looking for interesting directions. Maybe one image has amazing lighting, but the robot design is weak. Maybe another has a great robot, but the composition is boring. I’m collecting raw materials.

3. Remix and Refine: Now I start creating new prompts based on what I’ve learned. I might take the seed from the image with the great lighting and combine it with a new description of the robot. I’ll try different camera angles (--ar 16:9 for a cinematic look, --ar 3:4 for a portrait). I’ll experiment with the --style raw parameter to get a more photographic look, or --stylize values to control how artistic the image is. This is a process of constant experimentation. I might create 10, 20, or even 50 different prompts, each one a slight variation on the last.

4. Curate with a Vengeance: This is the most important part. Once I have a pool of 50-100 images, I start the curation process. I’m looking for the one image that not only looks amazing but also perfectly nails the objective I defined in Stage 1. I’m not just picking the “prettiest” image. I’m picking the most effective one. I’ll often pull my top 5 candidates into a separate folder and walk away for an hour. When I come back, the winner is usually obvious. It’s the one that makes me feel something. It’s the one that tells a story.

This process is not fast. It can take hours. But this is what it takes to create work that stands out. You’re not just a prompter. You’re a curator. You’re an editor. You’re a creative director. You’re guiding the AI to a specific destination, not just letting it wander aimlessly.

Stage 4: Post-Production & Polish

This is the final 10%, and it’s what separates the professional from the amateur. Almost no image that comes out of Midjourney is 100% perfect. There will be weird artifacts, strange details, or just things that could be better. The pros don’t just accept the AI’s output as final. They take it and they make it their own.

My post-production process is usually pretty simple. I’m not a professional photo editor, but I know a few basic moves that can make a huge difference. I usually use Photoshop, but you can use any image editor you’re comfortable with.

Here are the things I’m looking for:

• Color & Contrast: I’ll almost always do a final color grade. I’ll adjust the curves to make the blacks blacker and the whites whiter. I’ll tweak the saturation to make the colors pop, or desaturate them to create a more muted, moody feel. This is your last chance to really nail the visual language you defined in Stage 2.

• Cleaning Up Artifacts: AI can be weird. Sometimes you’ll get an image with six fingers on a hand, or a tree that’s growing into a building. I’ll go in and clone stamp these out. It’s tedious work, but it’s essential. Nothing screams “lazy AI art” like a glaring anatomical error.

• Compositing: Sometimes, one image isn’t enough. I might love the robot from one generation, but the background from another. I’ll take both images into Photoshop and composite them together. This gives you an incredible amount of creative control. You’re no longer limited to what the AI can generate in a single pass. You’re using the AI to create raw materials, and then you’re assembling them into a final vision.

• Adding Text & Graphics: If the image is for a social media post or a blog header, I’ll add text and other graphic elements in this stage. This is another opportunity to reinforce your brand and make the image more informative.

This final stage is what really makes the image yours. You’re not just a prompter anymore. You’re a creator. You’re taking the AI’s output and you’re imbuing it with your own taste, your own skill, and your own point of view. It’s the difference between a generic AI image and a piece of art that people will remember.

This Is Not the End of Creativity. It’s the Beginning.

I hear a lot of fear out there. Fear that AI is going to replace artists, designers, and creators. I see it differently. I’ve been investing in AI for a long time, and I’ve seen what it can do. It’s not a replacement for creativity. It’s an amplifier. It’s a tool that can help you bring your ideas to life faster and more powerfully than ever before.

Think about the invention of the camera. Did it replace painters? No. It created a whole new art form. It freed painters from the need to create realistic depictions of the world and allowed them to explore new forms of expression. The same thing is happening with AI. It’s freeing us from the technical grunt work of creation and allowing us to focus on what really matters: the idea, the story, the emotion.

The framework I’ve outlined here isn’t just about creating pretty pictures. It’s about a new way of working. It’s about being a creative director, not just a technician. It’s about having a vision and using the most powerful tools in the world to bring that vision to life. The people who master this new way of working are the ones who are going to create the iconic images of our time.

I’ve used this exact system to create images that have been seen by millions of people. The founders of my portfolio companies are using it to build stronger brands and more effective marketing campaigns. It works. But it’s not a magic bullet. It takes discipline, it takes taste, and it takes a willingness to put in the work. You have to be willing to go beyond the first, easy answer.

So, the next time you open Midjourney, don’t just ask it for a picture. Give it a job to do. Give it a story to tell. Give it a feeling to evoke. Don’t settle for “good enough.” Push the tool, and push yourself. The results will speak for themselves. You might not get to ten million views overnight. But you’ll be on your way to creating work that matters. And in the end, that’s the only metric that really counts.

Frequently Asked Questions

What tools do I need to get started?

Start with the basics. You don't need expensive software or fancy tools. A spreadsheet, a note-taking app, and direct access to your customers will get you further than any enterprise platform. Add tools only when you hit a specific bottleneck.

What are the most common mistakes when using midjourney to create results that got 10m+ views?

The biggest mistake I see is overcomplicating things early on. Start with the simplest version that works, get real feedback, and iterate from there. Another common trap is copying what worked for someone else without understanding the context behind their decisions.

How long does it take to use midjourney to create results that got 10m+ views?

The timeline varies depending on your starting point and resources. For most founders, expect 2-4 weeks for initial setup and 2-3 months to see meaningful results. I've seen teams move faster when they focus on one thing at a time rather than trying to do everything at once.