In this second episode of Chasing Creativity with Chris, I’ll share some new portrait sketches, a DIY coffee cabinet project and my continuing experiments with vibe coding art that you can try on your own. For my IT friends, click here to jump straight to the vibe coding section.
Portrait practice makes perfect?
I’ll let you in on a little secret: I almost never create a “finished” portrait. While my shelves are full of art books and my browser history is full of tutorials, I know that the real change only happens with practice. You have to put in the miles. For me, that means focusing on the journey, not the destination. My practice is a constant flow of sketches, not a series of masterpieces.
When it’s time to just play and loosen up, I keep things simple. I turn to vine charcoal and cheap newsprint. There’s no pressure. The materials are forgiving, and the goal isn’t to create a perfect drawing, but to make quick, honest marks and learn something in the process.
Sometimes, I’ll try different methods. In this case, I wanted to try a more structured or angular approach to the drawing as it results in a completely different look to the portrait.
In some cases, I’ll see something that will give me an idea for a concept I might explore further later. In this sketch, I wanted to explore blending the flow of ink or water through a portrait. This one was done without any reference. I’m finding the more I practice, the easier it gets to draw faces from imagination only.
There’s a peculiar anxiety that comes with a fresh sketchbook, isn’t there? A fear of making the first mark, of spoiling the perfection of all that blank paper. I found an unexpected antidote to this feeling while doing a bit of house cleaning.
As I was sorting through old books to get rid of, it occurred to me: Why was I seeing these as disposable items when they could be repurposed canvases? This wasn’t about ruining a pristine page; it was about giving an old one a new life. Many of the books had paper with a wonderful tooth, the pages were obviously not blank, and my new art supply was “free.”
This shift in perspective led me to start a new series of portraits directly in my old Strengths Finder book. The underlying text becomes part of the art, creating a side effect that adds a surprisingly compelling layer to the sketches.
Coffee Bar Cabinet Project
While drawing and coding satisfy one part of my creative brain, there’s a different kind of fulfillment that comes from building something tangible. The latest mission: a custom coffee bar for my daughter’s new home.
Maybe it’s the retired architect in me, but every project begins long before I touch a tool. It starts with a design. After referencing a few cabinets she admired, I did what felt most natural—I started sketching. Those initial drawings evolved into a detailed, to-scale plan that mapped out every component. It’s this methodical approach that translates an idea into a physical reality. While the final cost for materials is higher than a big-box store alternative, there’s simply no comparison. This is solid wood, fully custom, and built with a story.
The next steps for me including pricing all of the materials and rechecking the design to ensure the total cost is within budget. I then get to apply my analytical skills to find ways to optimize the materials and potentially use leftover materials from prior projects.
Once the materials are on-hand, we get to the fun part which is the build. I’ve included a few of the in-process stages below highlighting that I’m by no means a pro at this but have enough tools and equipment to get the job done. Usually each project “requires” me to buy a new tool, but nowadays I seem to always have just enough to get by. For this project the portable table saw and automatic brad nailer were the most important as they make it far easier to get square cuts and assemble quickly.
The final product turned out quite nice, especially with the addition of a barn door instead of the more traditional hinged door.
P.S. if you like espresso, I highly recommend La Terza and Coffee Emporium. 🙂
Vibe coding art
If you’ve read some of my prior posts you know that I enjoy figuring how to leverage AI to not only enhance my creativity but also to continue to keep up with the rapid pace of AI through practice. Based on everything I watch and read, and from what I see with my own eyes, AI continues to advance at an almost exponential pace. I’ll share more about this and Dario Amodei’s recent warning in a separate post.
But, to help make the point, I wanted to see if instead of simply helping me write code which creates visual art, I wanted to see if they could create every line of code from concept to final product. As you can already guess, the experiment was successful and I’ll share a few of them below and allow you to play with them yourself.
In the past, I’ve been using ChatGPT, but with Google’s recent new model announcements and excellent integration with other Google products, I switched my subscription to Google Gemini Pro for at least the next two months (1st month is free).
Image Art
“I’d like to create p5.js code which will manipulate a given image by first dividing it into a grid, then plotting each section of the image into 3d space such that each section has a random z position such that I can then use the mouse to rotate the overall grid in 3d space. This should create a dramatic artistic affect on the original image.”
With only that prompt, Gemini got started thinking and then created perfect code (no errors and did what I asked). I’ve included a snippet of the its thought process below. What you should notice is that it is very similar, if not better, than what a human would do.
What’s nice about Gemini is that instead of just sharing the code, it will automatically put it into Canvas where you can see the code being written and, even better, provides a Preview tab to see the resulting product. This provides a nice working space, or canvas, to iterate with Gemini to get more creative and enhance the initial code. By the way, it took a total of 30 seconds to think, create the code, and produce the output.
Some of the enhancements included adding sliders to adjust parameters and a button to allow uploading your own image. Below are 4 examples using 3 Midjourney created images and one of my previous code generated art images as the base.
It’s a bit difficult to see the full effect from a static image, so I’ve also provided a working example here. It’s best used on a PC or Mac so you can more easily upload an image and use the mouse to move around the image and change the sliders. Hold down left mouse button to rotate and right mouse button to drag. Hit the ‘S’ key to save you creation.
The ability to reason and ultimately generate working code which more or less delivers exactly what I had in mind is a major step forward and should be considered a bit scary. It gets even more scary when you continue the conversation with Gemini. There were a few cases where it didn’t do exactly what I wanted so I had to push back and the way Gemini responded back and provided accurate corrections was a bit disconcerting. It’s like working with a live javascript expert who is also conversational.
Because I can continue to iterate and it’s still a fully functional chatbot with broad-based knowledge, I can then ask it get creative and suggest ways to expand beyond my idea to generate even better and more creative outputs.
Substrate
Since this was working so well, I thought I would give Gemini a shot at creating something similar to a generative art project titled “Substrate” by James Tarbell.
I’ve always been intrigued by his work and how to generate it from code. I tried several times but nothing ever panned out. With that as the background, I wasn’t pleasantly surprised that Gemini was able to generate the foundation for this, with a few suggested variations, in a few minutes. I’ve included 4 examples below. You’ll note that I added the ability to create curved paths instead of just 45 and 90 degree angles. It’s also not as high resolution, nor as artistic. Yet. Look for a future update.
As before, you can try it out for yourself here. Use the sliders to adjust the parameters and click on the image to regenerate. ‘S’ to save your creation.
This is just a small sample of what’s possible with AI and LLM’s these days and things just continue to improve. If you aren’t already using these tools, you should at least get familiar with them as they can not only help you get more efficient but improve your ability to create and innovate. More to come on this in a future post.
Links to Working Examples
- Image fragmenter – https://cmartcreations.com/code/art/imagegrid3d.html
- “Substrate” – https://cmartcreations.com/code/art/substrate.html
So there you have it, another episode of random experiments in creativity which I hope provides some level of inspiration or questions.
As always, would love to get your comments below or email me directly at info@cmartcreations.com
