Milestone

Redefining Originality in the Age of AI

An experimental project exploring what it means to create original art when AI can generate images from billions of pictures. Trained entirely on personal works and public domain images, this is a search for authentic self-expression in a world of infinite replication.

View Gallery Read the Story

The Story Behind Milestone

I'm Jason Juan—an oil painter, mathematician, business entrepreneur, tech and entertainment industry leader, and award-winning artist. My career has spanned the intersection of art, technology, and innovation, always searching for new ways to merge creative vision with technical possibility.

Milestone represents a deeply personal experiment born from this unique perspective. With a foundation in mathematics and decades of experience bridging art and technology, I set out to answer a fundamental question: what does originality mean when AI can generate images trained on billions of pictures from the internet?

The answer, I believed, lay in creating something entirely personal—a model trained exclusively on my own photographs, drawings, paintings, and carefully selected public domain images. No borrowed aesthetics. No statistical averages of internet trends. Just my vision, through 13,304 images with 95.9% from just 7 days travel photos.

The name "Milestone" came from my last conversation with Gary Faigin on November 25, 2024, when I presented this project to him. Gary immediately grasped its significance. "Milestone," he said. "That's a good name for it."

Gary Faigin passed away on September 6, 2025, just a few weeks ago. As the founder of Gage Academy of Art in Seattle, Gary was more than an amazing artist—he was one of the greatest art educators and a true visionary. This project is dedicated to his memory.

In 2010, Gary contacted me to develop Gage Academy's first digital figure painting classes. At a time when many in the traditional art world were skeptical of digital tools, Gary expressed that digital painting was a new type of art, even though it was just the beginning. He understood what I had learned through my journey across mathematics, business, and art: tools don't define art—vision does.

Gary saw what Milestone represented: not just a technical achievement, but a philosophical statement about originality, authenticity, and the future of artistic expression. In a world where AI models are trained on billions of images scraped from the internet, this was a different path—a personal one. A path that honors both the artist's unique perspective and the legacy of those who believed in the evolution of art.

The Technical Foundation

13,304
Original Images
550M
Model Parameters
10 Days
Training Duration
768×768
Resolution

Dataset Composition

Personal unfiltered travel photos (7 days) 95.9% (12,765 images)
Personal artwork (drawings & paintings) 2.7% (355 images)
Public domain images 1.4% (184 images)

Trained from scratch on a single NVIDIA RTX 4090 GPU. A custom-designed diffusion model architecture built to prove that originality doesn't require billions of images—just a focused, personal vision.

Why This Matters

Since 2022, when Stable Diffusion 1.x was trained on LAION 5B—a dataset containing 5 billion images scraped from the entire internet—image generation has fundamentally changed. Anyone can now create stunning visuals in seconds. But in this new world of infinite generation, what does originality mean?

As AI increasingly blurs the lines between creation and replication, the question of originality requires a new definition.

Milestone is an experiment in attempting to answer that question. By training a model solely on personal works and public domain images, this project demonstrates that AI can generate images reflecting a unique artistic vision—not a statistical average of the internet.

The results speak for themselves. When posted on Reddit, most people enjoyed it. One skeptical user believed the model must have been fine-tuned or trained on more than 13,000 images—they couldn't believe it was built from scratch with such a small, personal dataset. I take this skepticism as the highest compliment. It means the results achieved what they set out to do.

Milestone isn't just a model name or a dataset—it's a philosophical statement. It's a small step, but a hopeful one, toward a future where AI can be a tool for authentic self-expression rather than merely replicating what already exists.

In an age of infinite replication, originality is not about having more data. It's about having a vision worth expressing.