I’m going to save you two years of your life.
That’s how long I’ve been obsessed with synthetic media. I’ve gone deep, invested a ton of time, and frankly, made a lot of mistakes. I’ve seen what works, what doesn’t, and what the so-called “experts” don’t tell you.
Most of what you read online is either surface-level, academic, or just plain wrong. It’s written by people who haven’t spent thousands of hours in the trenches, actually trying to get professional results. They haven’t built and sold companies, or invested in over 200 startups like I have. They don’t have the skin in the game.
I do. And I’m here to give you the three most critical, non-obvious lessons I’ve learned. This isn’t another “intro to AI” post. This is the real stuff, the hard-won knowledge that will accelerate your learning curve and save you from costly mistakes.
Lesson 1: Your Taste is the Most Important Skill
This might sound counterintuitive. In a world of complex models and technical jargon, how can something as subjective as “taste” be the most important skill?
Because it is.
I’ve seen people with immense technical knowledge produce mediocre, soulless synthetic media. And I’ve seen artists with a great eye for composition and light create stunning visuals with basic tools.
The models are getting better every day. The technical barriers are coming down. What will always be in short supply is good taste. The ability to know what looks good, what tells a story, and what evokes an emotion.
Think about it. When you’re generating an image, you’re not just typing in a prompt. You’re a director. You’re choosing the lighting, the composition, the mood. You’re making a thousand tiny decisions that all add up to the final image.
And those decisions are guided by your taste.
I learned this the hard way. When I first started with Stable Diffusion, I was obsessed with the technical side. I spent weeks learning about samplers, CFG scale, and all the other arcane settings. I could generate technically perfect images, but they were boring. They lacked a point of view.
It wasn’t until I started focusing on the art direction that my images started to improve. I started thinking about light, color, and composition. I started to develop my own style.
So, how do you develop your taste?
- Study the masters. Look at the work of great photographers, painters, and filmmakers. Analyze their use of light, color, and composition.
- Create, create, create. The more you create, the more you’ll develop your own eye. Don’t be afraid to experiment and make mistakes.
- Get feedback. Show your work to people you trust and get their honest feedback. This is how you’ll learn and grow.
Lesson 2: The “Last Mile” is Everything
Getting a synthetic media output that’s 90% of the way there is easy. The tools are good enough now that anyone can generate something that looks pretty good.
But that last 10%? That’s where the magic happens. That’s what separates the amateurs from the pros.
I call this the “last mile” problem. And it’s the biggest challenge in synthetic media today.
What do I mean by the “last mile”? I’m talking about all the little details that make an image or video look real. The subtle imperfections, the way light reflects off a surface, the way a person’s hair moves in the wind.
These are the things that are incredibly difficult to get right with current models. And they’re the things that will make or break your final output.
I’ve spent countless hours trying to fix these “last mile” problems. I’ve tried everything from inpainting and outpainting to using multiple models and complex post-processing workflows.
And you know what I’ve learned? There’s no magic bullet.
It’s a grind. It’s a lot of trial and error. But it’s also where you can really add value.
If you can master the “last mile,” you’ll be able to create synthetic media that is indistinguishable from reality. You’ll be able to create things that no one else can.
Here are a few tips for tackling the “last mile”:
- Don’t be afraid to get your hands dirty. You’re going to have to do some manual work. Whether it’s inpainting, rotoscoping, or color grading, you’re going to have to get in there and fix things by hand.
- Use multiple tools. Don’t rely on a single model or tool. Use a combination of tools to get the job done. For example, you might use one model to generate the initial image, another to fix the hands, and a third to do the final color grade.
- Think like a VFX artist. Visual effects artists have been dealing with the “last mile” problem for decades. Learn from their techniques and workflows.
Lesson 3: The Future is Multimodal
Right now, we have models that are really good at generating images. We have models that are really good at generating text. And we have models that are really good at generating audio.
But the real magic will happen when we can combine these models into a single, multimodal system.
Imagine being able to write a script, and then have an AI generate a complete video, with realistic actors, dialogue, and sound effects.
That’s the future of synthetic media. And it’s not as far off as you might think.
I’m already seeing the first signs of this multimodal future. I’m seeing tools that can generate a 3D model from a single image. I’m seeing tools that can generate a video from a text prompt.
These tools are still in their early stages, but they’re improving at an exponential rate.
So, what does this mean for you?
It means that you need to start thinking multimodally. You need to start thinking about how you can combine different types of media to create new and interesting experiences.
Don’t just think about generating images. Think about generating images, text, and audio. Think about how you can use these different media to tell a story.
This is where the real opportunities are. The people who can master multimodal creation will be the ones who shape the future of synthetic media.
The Takeaway
I’ve just given you the three most important lessons I’ve learned after two years of deep immersion in synthetic media. These aren’t the things you’ll read in a textbook. This is the real, hard-won knowledge that will give you an edge.
So, what are you waiting for? Go out there and start creating. The future is in your hands.
Frequently Asked Questions
Can I implement all of these at once?
I'd strongly recommend against it. Pick the 2-3 items that resonate most with your current situation and focus there. Trying to do everything simultaneously is a recipe for doing nothing well.
How were these items selected?
Each item on this list comes from direct experience, either from building my own companies or from patterns I've observed across the 200+ startups I've invested in. I prioritize practical, actionable items over theoretical concepts.
Which item on this list has the highest impact?
It depends on your stage and context, but in my experience, the items near the top of the list tend to have the broadest applicability. That said, sometimes the less obvious items create the biggest breakthroughs for specific situations.
How do I know which items apply to my situation?
Start by honestly assessing where your biggest bottleneck is right now. The items that address that specific constraint will give you the highest return on your time and energy.