Unraveling the Code Part 3: Latent Space Exploration and Fruits of Labor

Between Part 2 of this blog series and now, I took the project from version 14 to 15. I promised myself I wouldn’t make any more changes to it until I share the fruits of my labor with the world, but my obsession got the best of me. 

Between version 12 and version 14, I made a noteworthy change to the project. I renamed it from ABSTRACT LANGUAGE to VISUAL ABSTRACTION. If you paid close attention to Part 1’s title, you might have anticipated this. With nearly half of my life spent in engineering (5 years in Mechanical Engineering during college and 11 years in Software Engineering in the corporate world), I tend to inadvertently give my creative projects academic-sounding names, sometimes even using titles that can be easily turned into acronyms. ABSTRACT LANGUAGE felt too much like a linguistics research paper and didn’t capture the creative aspect of this work, so I opted for VISUAL ABSTRACTION to evoke a more artistic flair.

Let’s dive into the core of this post and discuss my reasons for incorporating AI into my work, as well as offering a glimpse into my workflow exploring the Latent Space. To begin, I’ve been exploring this field much longer than most people currently producing “AI Arts.” I’m well aware of the concerns surrounding AI, and I share quite a few of them myself. I have no fewer than 100 from my photography work in the LAION dataset that is used to train the AI models used in Stable Diffusion. I also have a solid understanding of how the diffusion process work to generate images. I chose to use AI in my workflow simply because it’s another tool that allows creatives to broaden their horizons. Ultimately, I was a full-fledged artist before AI and would have continued to be one without it. I have more thoughts on AI in general, but I’ll save those for future posts.

Every project or image I’ve created with the assistance of AI so far involves at least one custom model that I’ve fine-tuned using my work or a base image from my archive. This will change soon however as I’m going to start an exploration using MidJourney… In this specific project, I used a combination of three style models, fine-tuned on numerous artistic endeavors I’ve undertaken over the years. Revisiting the farmer and garden analogy from Part 2, this is where some of my less successful crops and livestock find new purpose as fertilizer. Below, you’ll find a collage showcasing some images from the dataset I used to fine-tune one of these three models, accompanied by details on each exploration.

Row 1: My handmade sculptures. Stay tuned for an upcoming post on the creative process behind these pieces.

Row 2: My early AI explorations using Disco Diffusion. These pieces inspired an oil painting I did that merges chaos and modern cubism. I’ll likely revisit this painting style in the future.

Row 3 and 5: Some of my initial digital line drawings based on my photography. A few of these pieces were created trips to Paris and London a few years back. Even back then, I felt these pieces would play a role in my artistic journey. Check out the caption for this Instagram post from 5 years ago.

Row 6: Exploring color studies and organic shapes. The 2nd and 4th images are digital paintings from the early pandemic days when I challenged myself to create daily. This led to discovering MagicaVoxel and transforming the pieces into 3D voxels.

Row 4: A progression of the works in Rows 3, 5, and 6. Intrigued by the amazing works some artists achieved using MagicaVoxel, I sought to bring something distinct to the table. This series, In Living Color, is also inspired by my photography.

Row 7: Yet another exploration using Disco Diffusion, where I became obsessed with the psychedelic futuristic visual effects, I could achieve with it.

If you’re curious about how to fine-tune an AI model based on your work or seamlessly integrate AI into your creative process, don’t hesitate to get in touch with me. I can provide you with some helpful online resources to guide you. If that’s not sufficient, feel free to inquire about my services. And yes, in case you were wondering, this was a little ad break right in the middle of the post.

Now, returning to our regularly scheduled content. By utilizing the style models and the generated art piece from version 15 of VISUAL ABSTRACTION, I’ve achieved my ultimate goal: cultivating a self-sustaining creative garden that enables me to produce artwork that’s not only unique and visually captivating but also carries literary significance. Without further ado, let’s have a look at the fruits of my labor:

Throughout this creative journey, “LOVE IS A DRUG” has consistently been the recurring text that serves as my primary test case. It’s quite fitting for this project, as it has truly been a labor of love. On the topic of LOVE, I used to joke about people who had “LIVE, LOVE, LAUGH” wall decor in their homes, as it seemed generic to me. However, as I’ve aged, I’ve come to realize that those words might genuinely hold meaning for the people who live in the home or make the space feel cozier. Nevertheless, I still believe it would be better to incorporate those words into a more visually appealing and unique piece of art. So, here’s a gif showcasing the UI as of version 15 and the word LOVE in various representations:


Here’s LOVE in a cozy home:

Here’s LOVE again in a colorful abstraction: 

Lastly, here is a representation of the Haiku from Part 2. Since by design a Haiku is composed of 3 lines, it is fitting to represent it as triptych. This also shows the flexibility of this project to fit the wildest dream of an interior designer, collector, or homeowner when it comes to decorating a place with unique art. 

So, what’s next for this project? I have a bunch of ideas lined up, but for the time being, I need to take a break from tinkering with the code. I’m also considering a few ambitious concepts involving other mediums, which will surely elevate the ecosystem created by this project to new heights. Keep an eye out for more updates on that front.

For now, I invite you all to savor the fruits of my labor. Please visit this page to see more art from this series. I’ll be doing periodic drops in my store for this project, giving everyone a chance to get their hands on some pieces. For a more personalized experience with this work, don’t hesitate to reach out to me directly.

Keep these important points in mind:

  • The artworks from this project are meant to be displayed as a diptych, at the very least.
  • Each generative piece from my algorithm comes with at least one related piece derived from it.
  • Every combination will be a unique limited edition (1 of 1) as physical art. If a collector prefers digital assets, the work can be minted for them.

Thank you for taking the time to read this blog and peek into the creative chaos in my mind.

Using Format